Explore: The Journal of Science and Healing
Volume 3, Issue 6 , Pages 561-572, November 2007

Rationale and Methods for a Trial Assessing Placebo, Echinacea, and Doctor-Patient Interaction in the Common Cold

  • Bruce Barrett, MD, PhD

      Affiliations

    • Department of Family Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, WI
    • Corresponding Author InformationCorresponding Author. Address: 777 S Mills, Madison WI 53715
    • ,
  • Dave Rakel, MD

      Affiliations

    • Department of Family Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, WI
    • ,
  • Betty Chewning, PhD

      Affiliations

    • School of Pharmacy, University of Wisconsin, Madison, WI
    • ,
  • Lucille Marchand, MD

      Affiliations

    • Department of Family Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, WI
    • ,
  • David Rabago, MD

      Affiliations

    • Department of Family Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, WI
    • ,
  • Roger Brown, PhD

      Affiliations

    • School of Nursing, University of Wisconsin, Madison, WI
    • ,
  • Jo Scheder, PhD

      Affiliations

    • Department of Human Development and Family Studies, University of Wisconsin, Madison, WI
    • ,
  • Raandi Schmidt, APNP

      Affiliations

    • Department of Family Medicine, School of Nursing, University of Wisconsin, Madison, WI
    • ,
  • James E. Gern, MD

      Affiliations

    • Department of Pediatrics, University of Wisconsin, Madison, WI
    • ,
  • Kerry Bone, Dip Phyto

      Affiliations

    • MediHerb, Warwick, Australia
    • School of Health, University of New England, Armidale, NSW Australia
    • ,
  • Gay Thomas, MA

      Affiliations

    • Department of Family Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, WI
    • ,
  • Shari Barlow

      Affiliations

    • Department of Family Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, WI
    • ,
  • James Bobula, PhD

      Affiliations

    • Department of Family Medicine, School of Medicine and Public Health, University of Wisconsin, Madison, WI

Article Outline

Background

Clinical medicine and healthcare policy are increasingly guided by randomized controlled trials, which in turn are dependent on the validity of placebo control. It is important to understand the effects of placebo control on outcome measurement, especially for assessment of symptoms and functional impairments where subjectivity, expectancy, and motivation may significantly impact outcome evaluation. This paper describes the rationale and methodology of a trial designed to evaluate placebo effects related to taking pills and to compare these with effects attributable to standard or enhanced (patient-oriented) doctor-patient interaction.

Design

This trial uses two-way factorial allocation to randomize people with new onset common cold in two directions: pill related and doctor related. In one direction, participants are randomized to (1) no pills, (2) blinded placebo, (3) blinded echinacea, or (4) unblinded open-label echinacea. In the other direction, participants are randomized to: (1) no doctor-patient interaction, (2) standard doctor-patient interaction, and (3) enhanced doctor-patient interaction. Enhanced interaction includes education, empathy, empowerment, positive prognosis, and connectedness. Area under the time severity curve is the primary outcome, with the Wisconsin Upper Respiratory Symptom Survey (WURSS-21) the measure of severity. A priori power studies called for a sample size of N = 720 trial finishers to detect 15% to 20% between-group differences in this outcome. Secondary outcomes include general health-related quality of life, perceived stress, interpersonal support, optimism, patient satisfaction, and positive and negative affectivity. Two biomarkers are also assessed: interleukin-8 (inflammatory cytokine) and neutrophil count from nasal wash.

Importance

This paper describes the rationale and methodology of a trial assessing placebo effects related to pills and to doctor-patient interaction. This is one of very few similar studies and is the first in the common cold. Data collected will also provide an excellent opportunity to investigate relationships among demographic (age, sex, education, income) and psychosocial (perceived stress, interpersonal support, optimism, affectivity) indicators in relation to common cold outcomes.

Key words: Clinical trials, common cold, communication, double blind method, physician-patient relations, placebos, placebo effect, respiratory infection, therapeutics, treatment outcome

 

Back to Article Outline

Introduction 

Clinical medicine and healthcare policy are increasingly guided by evidence from randomized controlled trials (RCTs).1, 2 The purpose of randomizing sufficient numbers of people into intervention and control groups is to evenly distribute all outcome-influencing factors and thus avoid or mitigate numerous potential biases.3, 4, 5, 6, 7, 8 Allocation concealment (masking, blinding) is also considered important, as it keeps the expectations and preferences of both researcher and subject from biasing outcome assessment.9, 10, 11, 12, 13 This is especially true for assessment of symptoms and functional impairments where subjectivity, expectancy, and motivation may significantly impact outcome evaluation. Nevertheless, although randomization to a blinded placebo control group is standard practice in medical research, there are substantive complexities that are too often overlooked. This paper describes the rationale and methodology of an RCT designed to evaluate placebo effects related to taking pills and to compare these with effects attributable to standard or enhanced (patient-oriented) doctor-patient interaction.

Conventional theory defines the average difference in outcome between placebo and intervention groups as the specific effect of that treatment. Placebo effects,14, 15, 16, 17, 18, 19 Hawthorne effects,20, 21, 22 doctor-patient interaction effects,23, 24, 25, 26, 27 and other factors that might influence health-related outcomes are considered nonspecific effects28, 29, 30 or context effects31, 32 and are controlled for by randomization and blinding. However, although it is implicitly acknowledged that nonspecific mechanisms contribute importantly to health-related outcomes, humanity’s vast medical research enterprise has focused on specific effects, leaving other health-influencing factors largely unexplored.

There are at least two fundamental problems encountered when attempting to extrapolate the results of RCTs to real life (ie, nonresearch settings). The first derives from the fact that the people who participate in trials are not randomly drawn from the larger population, and hence are not likely to be representative of it. Similar problems would be encountered by researchers attempting to predict a national election from the results of a survey of volunteers in a single voting district. Potential solutions to this problem of selection bias include directed attempts at representative samples, such as large multicenter trials, and/or comparative analysis of multiple single-center trials.

The second problem stems from issues related to blinding and the choice of comparison group. In real life, people choose to take or not take a treatment and are aware of both their choice and the intervention. In conventional research context, people are randomized to blinded treatment versus blinded placebo and hence are aware of a 50% chance of a possibly active intervention. If conscious awareness and expectation influence outcomes, then extrapolation of research results to nonresearch settings may be erroneous. The purpose of using the blinded placebo is to control for expectancy and other context effects. However, if these nonspecific mechanisms influence outcomes, then specific effect estimates may be inaccurate predictors of real-life treatment effects. Partial solutions to this problem may include additional comparison groups, such as no-pill groups, open-label groups, and/or head-to-head trials comparing active interventions.

There is very little direct evidence regarding the effects of doctor-patient interaction on clinical outcomes. Two reviews and one trial will be summarized. A systematic review by Stewart27 reported that 16 of 21 studies reviewed reported positive health outcomes. Benefits ranged from decreased anxiety, distress, and pain, to improvements in blood pressure, mobility, and general health-related quality of life. Effect sizes ranged from 20% to 40% for anxiety scores, to a three-fold difference in resolution of headache.27 Di Blasi et al31 published a systematic review of context effects on health outcomes. They reviewed 25 randomized trials and reported that there were more positive than negative findings, with effect sizes ranging from small to large. They did note that “one relatively consistent finding is that physicians who adopt a warm, friendly, and reassuring manner are more effective than those who keep consultations formal and do not offer reassurance.”31

A highly cited RCT by Thomas33 randomized 200 consecutive patients “in whom no definite diagnosis could be made” in two directions, using a balanced two-way factorial design. In one direction, patients received either open-label placebo (3-mg thiamine) or no prescription (no pills). In the other direction, participants received either a positive consultation in which “the patient was given a firm diagnosis and told confidently that he would be better in a few days,” or a nonpositive consultation in which the doctor said “I cannot be certain what is the matter with you.” For those receiving the prescription (placebo), positive words were said about the pills, whereas in the nonpositive group the doctor said, “I am not sure that the treatment I am going to give you will have an effect.” The primary outcome was whether or not the patient reported resolution of illness at two weeks follow-up. In this study, the thiamine prescription did not appear to influence illness resolution (53% vs 50%; P = .5). Positive consultation, however, led to 64% reported recovery, versus 39% in the nonpositive group (25% absolute effect size; P < .01; number needed to treat is four). To our best knowledge, this is the only RCT that has tested both pill placebo effects and doctor-patient effects in the same trial.

This paper serves as a methodological reference point for an ongoing RCT entitled “Placebo: Physician or Pill: A Randomized Controlled Trial in a Common Cold Model.” We also call this the PEP trial, because the main interventions are physician, echinacea, and placebo. This trial arose in response to a Request for Applications to the National Center for Complementary and Alternative Medicine at the U. S. National Institutes of Health entitled “The Placebo Effect in Clinical Practice.” A literature review and conceptual framework paper has been published.19

Back to Article Outline

Methods 

This trial was built on a foundation of previous work. Several years ago we carried out a conventional RCT testing echinacea versus placebo as treatment for the common cold.34 When designing that trial, we discovered that there was no well-designed, validated instrument for measuring common cold outcomes. To fill that gap, we developed and tested the Wisconsin Upper Respiratory Symptom Survey (WURSS), which is now supported by reliability, responsiveness, convergence, and other validity data.35, 36, 37, 38 During early stages of that work, we realized that the common cold was an excellent venue for investigating placebo effects because: (1) nearly everyone experiences colds, hence there is a large and diverse source of research participants and (2) there are no proven effective cold treatments, hence, few ethical concerns for trials requiring no-treatment groups. Thus, when the National Center for Complementary and Alternative Medicine unveiled its placebo-effect RFA, we were ready with an illness model, a validated outcomes instrument, and a focused set of placebo-related research questions.

The PEP trial was designed to test both pill-related and doctor-patient interaction–related placebo effects. The two-way factorial design allows independent randomization to (1) no doctor-patient interaction, (2) standard doctor-patient interaction, and (3) enhanced doctor-patient interaction. Standard interactions include the basic elements of a medical encounter (history of present illness, relevant past medical history, focused physical exam, diagnosis, and plan.) Enhanced visits include these elements, but also aim for five patient-centered domains: education, empathy, empowerment, positive prognosis, and connectedness (described in more detail below).

The same participants are also randomized to one of four pill-related conditions: (1) no pills, (2) blinded placebo, (3) blinded echinacea, and (4) unblinded open-label echinacea. Echinacea was selected as the active treatment because of the popularity of this herbal medicine and because of the suggestive but controversial evidence of its efficacy.39, 40, 41, 42, 43, 44 Placebo and echinacea pills were manufactured as identical-appearing coated tablets by MediHerb (Warwick, Queensland, Australia). MediHerb substantively contributed to the selection, manufacturing, and phytochemical evaluation of the echinacea product (and placebo), but was otherwise not involved with the design or conduct of the study.

The area under the duration severity curve was selected as the primary outcome, using the WURSS-21 instrument as the measure of severity.36, 37, 38, 45 Using variability estimates from previously collected data, power studies suggested the need for approximately 180 trial finishers per group to have 80% power to detect a 20% difference in this measure. Table 1 displays comparison group sizes in a schematic of the basic two-way factorial structure. While the trial was designed to have adequate power to detect differences among columns (pill effects) and to detect differences among rows (doctor-patient interaction effects), the power to assess differences or interactions between these two different sets of factors (rows vs columns, cell-level effects) is possible but not guaranteed.

Table 1. Two-Way Factorial Design of the PEP Triala
Allocation GroupsNo Pills, UnblindedPlacebo, BlindedEchinacea, BlindedEchinacea, UnblindedNo.
Novisit60606060240
Standard60606060240
Enhanced60606060240
Total180180180180720

aPEP trial (physician, echinacea, and placebo) is a shortened version of the trial title “Placebo: Physician or Pill: A Randomized Controlled Trial in a Common Cold Model.”

Participants 

The PEP trial recruits participants from the general population of Dane County, Wisconsin, with two enrollment centers: the University of Wisconsin, Department of Family Medicine attached to St. Mary’s Hospital, Madison, and the University of Wisconsin Health Family Medicine Clinic in Verona, 10 miles south of Madison. Potential participants must answer yes to either: “Do you think that you have a cold?” or “Do you think you are coming down with a cold?” They must report at least one of four named cold symptoms (or synonyms): nasal discharge (runny nose), nasal obstruction (nasal congestion, stopped up nose, stuffiness), sneezing, and sore throat (raw throat, scratchy throat). None of these may have arisen more than 36 hours before enrollment. Participants must have a total Jackson score46, 47, 48 of two or higher at enrollment, calculated by summing eight symptom scores, where zero is absent, one is mild, two is moderate, and three is severe. The eight symptoms include the four named above, plus headache, malaise, chilliness, and cough. These criteria were chosen to include people who would be diagnosed with viral upper respiratory infection (URI), or the common cold. Allergic disease is excluded by additional questioning if sneezing or itching of the eyes or nose is present. People with history of asthma are excluded if they have cough, wheezing, or shortness of breath at time of enrollment. Diagnostic accuracy is strengthened by the fact that two thirds of participants are examined by a physician or a licensed nurse practitioner.

Adults and minors aged 12 to 17 are potentially eligible. Minors must be enrolled in school in 6th grade or higher and must have written parental permission. Pregnancy is excluded by interview. Additional exclusion criteria include current use of antibiotics, antivirals or nasal steroids, or anticipated need for decongestants, antihistamines, or other symptom-relieving medications. People with autoimmune and immune deficiency disease (HIV infection, lupus, rheumatoid arthritis) are also excluded, as are people with known allergy to Echinacea or other tablet ingredients.

Interventions 

Interventions are defined in contrast to no pills and to no doctor-patient interaction. Therefore, blinded placebo, blinded echinacea, and open-label echinacea are the three pill-based interventions, and standard and enhanced doctor-patient interaction are the two clinical interventions.

Echinacea Tablets 

The echinacea tablets were produced by MediHerb (Warwick, Queensland, Australia) under the direction of K.D.B.P., and were designed to include all suspected active ingredients. These include the caffeic acid derivatives (caftaric acid, cichoric acid, echinacoside, chlorogenic acid, cynarin) and the lipophilic alkamides, including isobutylamides. Each echinacea tablet contains the equivalent of 1,275 mg of Echinacea root, as follows: (1) Echinacea purpurea, 675-mg root yields 112.5-mg dried extract, standardized to 2.1-mg alkamides; and (2) Echinacea angustifolia, 600-mg root yields 150-mg dried extract, standardized to 2.0-mg alkamides. Thus, each tablet is standardized to contain approximately 4.1-mg alkamides. Baseline phytochemistry assays done by MediHerb and by ChromaDex Inc (Clearwater, Fla) agreed that each tablet contained approximately 13.1-mg phenolic acids, composed of 1.8- to 2.1-mg caftaric acid, 6.3- to 8.9-mg cichoric acid, 3.8- to 4.3-mg echinacoside, and approximately 0.4-mg chlorogenic acid and 0.8-mg cynarin. MediHerb reported 4.5-mg and ChromaDex reported 12.4-mg total isobutylamides. Discrepancies are likely due to both analytic method and reference standards. Both labs found that 2,4-diene isobutylamides were the most prominent alkamides.

Standard tabletting excipients (calcium phosphate, cellulose, silica, sodium starch glycolate, hypromellose, and magnesium stearate) are added to produce a capsule-shaped tablet that is then coated to enhance blinding. The coated placebo tablets are identical in appearance and manufactured with the same process, with extra tabletting excipient ingredients instead of Echinacea.

Dosing and Adherence 

Echinacea is thought to work by stimulating the immune response to infection, and hence, should be begun as soon as possible. For this study, the first dose of two tablets occurs at enrollment (within 36 hours of first symptom), with three more doses of two tablets within 24 hours of enrollment. For the next four days, participants take one tablet four times daily. The treatment course is limited to five days because possible benefits of immune stimulation would decline once viral replication has diminished. Adherence is monitored by asking, “Did you take your pill treatment today?” and “How many pills did you take today?” during daily telephone monitoring calls. Also, participants bring their pill bottles to the exit interview, and any left over pills are counted.

Standard doctor-patient interactions are designed to meet ethical and customary standards of medical practice and are characterized by telling the patient “You have a cold, there’s not much you can do, as there is little evidence of benefit for any particular treatment.” Standard interaction attributes include:

•history of present illness (severity and duration of symptoms; treatments tried and perceived effectiveness)

•past medical history (drug sensitivities; history of allergy, asthma)

•physical exam (brief symptom-directed exam)

•diagnosis (usually “you have a common cold”)

•a plan (the clinician asks the participant to follow the instructions provided by the enrolling research assistant but provides little if any additional advice on self-treatment of colds)

The clinician is not informed—and does not ask—about pill-related group assignment. If the participant spontaneously divulges that he/she has received open-labeled echinacea, blinded pills, or no pill treatment, the clinician remains noncommittal about implications or possible effects of that allocation.

Enhanced doctor-patient interactions are defined as patient-oriented, participatory, and positive, with five attributes in addition to those in standard visits: education, empathy, empowerment, positive prognosis, and connectedness. These attributes (domains) were chosen because of their prominence in the patient-doctor interaction literature.27, 31, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67 We chose general domains rather than specific items or scripts because of the unpredictable nature of interpersonal interaction and the need for flexibility when responding to patients’ words and actions. Further rationale and description follows.

Education 

Authors such as Roter and Hall68 and Stewart27 have highlighted education as a factor associated with patient satisfaction and other outcomes. In our study, clinicians provide educational information regarding URI. Specific comments depend on the nature of a given interaction, but could be “A common cold is a viral infection of the upper respiratory tract. The length of illness is usually a week or less. The likely cause is a rhinovirus, which infects the inside lining of the nose and throat. The infection may cause symptoms, but is not dangerous.” In enhanced but not standard visits, clinicians provide written instructions on self-care of colds, including rest, stress avoidance, hand washing, fluids, and hot baths. Personalized comments (eg, “It’s fine to take your chamomile tea” or “Yes, it’s good to exercise, but try not to overdue it”) are handwritten on the information sheets.

Positive prognosis 

A number of authors have identified positive talk or positive attitude as an important component of quality clinical interactions.68, 69, 70, 71, 72, 73, 74, 75 Thomas’ trial,33 with its 25% effect size, provides some evidence of the power of a positive prognosis. In our study, enhanced clinical interactions include comments such as “You are going to be all right. Your cold is likely to resolve in the next few days. Generally, colds last only six days or so. Symptoms tend to be worse in the first few days. Your symptoms should be resolving soon.” For enhanced interactions, clinicians become aware of pill-related group assignment when opening the randomization envelope. Positive comments are tailored to the allocation status (eg, “Great! You got the real echinacea” [open-label], or “There’s a very good chance you got the real echinacea [blinded pills], or “At least you won’t have to bother taking pills” [no pill group]).

Empathy 

In a systematic review of 25 randomized trials, Di Blasi et al31 reported that a warm, friendly, and reassuring manner was associated with positive outcomes. In our study, physicians performing enhanced visits make extra effort to provide kind, empathetic care. Depending on the specifics of a participant’s concerns, the study physician may say something like “It’s normal to be worried about XYZ (eg, ear or sinus infection)” or “Yes, a cold can really take your energy away.” Attentive listening with empathetic facial expression is encouraged.

Empowerment 

Numerous authors have noted the association between various indicators of health and a sense of personal autonomy or empowerment.73, 76, 77, 78, 79 The patient-doctor interaction literature suggests that physicians can enhance a patient’s sense of autonomy and confidence through empowering talk.60, 75, 80 In our study, enhanced visits include comments such as “You can really make a difference in your cold by taking care of yourself.”

Connectedness 

Suchman and Mathews81 and Mathews and Suchman,82 among others,52, 59, 70, 83 have suggested that the degree of connectedness between doctor and patient is a predictor of health-related outcomes. During enhanced visits, study physicians seek to connect with their patients through eye contact, handshake greeting, humor when appropriate, and patient-oriented social and interactive discussion.

Verification and Evaluation of Standard and Enhanced Visits 

With permission, interactions are videotaped. A subsample will be reviewed with the 10 domains above (five standard and five enhanced), coded for presence or absence of these domains and rated for quality. Directly after doctor-patient interactions, participants fill out Mercer’s 10-item Consultation and Relational Empathy scale.84, 85, 86 Two additional items assessing liking87 and connectedness ask, “How much did you like this doctor?” and “How connected did you feel to him/her?” The answers to each question are scored: one is very little, two is not very much, three is somewhat, four is quite a lot, and five is very much.

Training of Study Clinicians 

Study clinicians (B.B., R.B., L.M., D.R., D.R., and R.S.) were trained over several months by a medical anthropologist (J.S.), who also has experience as a film and play director. Training included cognitive and behavioral components. Performance technique was refined using role-play in standard and enhanced modes, first with each other, then with mock patients, then finally with real patients presenting with cold symptoms. For final testing, each clinician enacted three standard and three enhanced interactions. All 10 domains were scored by a blinded videotape reviewer, who verified that the five enhanced domains were represented in the enhanced but not the standard interactions.

Back to Article Outline

Results 

Primary Outcome: Severity Days 

The primary outcome is defined as area under the time severity curve, a quantity which we call severity days. Severity is self-assessed twice daily using the Wisconsin Upper Respiratory Symptom Survey (WURSS-21), an illness-specific quality of life instrument.36, 37, 38, 45 Calculation of the variable for the area under the curve will follow the trapezoidal rule, as discussed below.

Secondary Outcomes 

Duration is defined as the time interval from the first reported symptom until the participant last replies yes to “Do you think you still have a cold?,” which must be followed by answering no for two consecutive days. The time of onset of the first cold-specific symptom is assessed retrospectively at enrollment. If a no is followed the next day by a yes, the cold is not over, but instead monitoring continues until the participant answers no two days in a row or reaches 14 days of observation. We decided to limit observation to a maximum of 14 days to avoid potential bias associated with very long illness or with people who are overly prone to rate themselves as sick.

Feeling Thermometer 

This single item general health marker that uses a 100-mm visual analog scale (VAS) to rate health on a scale where worst imaginable health condition is rated zero to best imaginable possible health is rated 100.88, 89, 90 The feeling thermometer is scored every day during participation.

General health-related quality of life 

The SF-36, also known as the Medical Outcomes Study 36-Item Short Form Health Survey, was developed as a generic health-related quality of life instrument and has been extensively validated.91, 92, 93, 94, 95 The SF-8 (24-hour recall) derives from the SF-36, with eight instead of 36 items, and a 24-hour instead of four-week recall period.96 The SF-8 is self-administered at enrollment, once each day throughout participation.

Expectation 

At enrollment, participants are asked to predict how severe they think their cold will be and how long they think it will last: “Compared to prior colds, this cold feels like it will be (1) much milder, (2) milder, (3) about the same, (4) more severe, and (5) much more severe; and “Compared to prior colds, this cold feels like it will last (1) much shorter than usual, (2) shorter than usual, (3) about the same as usual, (4) somewhat longer than usual, and (5) much longer than usual.”

Belief in echinacea 

At enrollment participants are asked, “Have you ever taken echinacea before?” and “How effective do you think that echinacea is?” A 100-mm VAS is marked at the top by extremely effective and at the bottom by totally ineffective. Participants are able to opt out by answering: (1) I have never heard of echinacea or (2) I have no opinion about whether or not echinacea is effective.

Belief in placebo effects 

At enrollment, participants mark a 100-mm VAS labeled “In general, how effective do you think that placebos are?” marked at the top by extremely effective and at the bottom by totally ineffective. A second question asks, “In general, how important is it that a person believes in his or her treatment?” A 100-mm VAS is marked extremely important at one end and completely unimportant at the other.

Perceived Stress Scale 

Stress has been linked to URI incidence, severity, and duration.97, 98, 99, 100 Cohen’s Perceived Stress Scale (PSS-4) has been used in a number of URI studies measuring biological as well as psychological variables.97, 101, 102, 103, 104 In this study, the PSS-4 is administered at enrollment, day 3, and at the exit interview. Additionally, a 100-mm VAS stress scale is scored once daily: “During the past 24 hours, how stressful has your life been?” One end is labeled extremely stressful and the other not at all stressful.

Interpersonal support and social support has been linked with many health conditions,79, 105, 106, 107 including respiratory infection.108 The PEP participants self-assess sociability and social support at enrollment, day 3, and the exit interview using Ryff’s nine-item Positive Relationships with Others scale.73, 109

Optimism: the Life Orientation Test 

A positive, optimistic attitude has been linked to a variety of health outcomes.71, 73, 110, 111, 112 The six-item Life-Orientation Test 113, 114 is administered at enrollment only.

Positive and negative affectivity 

Both positive and negative affectivity have been shown to influence the reporting of symptoms and dysfunctions in the common cold.115, 116, 117 The Profile of Mood States 18-item questionnaire allows participants to rate themselves by using adjectives such as happy, cheerful, energetic, sluggish, on edge, relaxed, tense, sad, and tired.118, 119, 120, 121, 122, 123 Profile of Mood States is self-administered at enrollment, on day 3, and at the exit interview.

Laboratory-Assessed Outcomes 

Biomarkers assayed from nasal wash at enrollment and day 3 follow-up include inflammatory cytokine interleukin-8 (IL-8) and neutrophil count. Five milliliters of buffered sterile saline is instilled in each nostril, and then collected in Petri dishes.

Neutrophil counts 

Counting neutrophils in nasal mucus is a relatively well-established indicator of nasal mucosal inflammation during URI.124, 125, 126, 127, 128 In our study, nasal wash is processed and stained using conventional methods. Neutrophils are counted within two hours of sample acquisition, and a fixed slide is prepared and stored as backup.

Interleukin-8 

Interleukin-8 is an inflammatory cytokine found in nasal secretions and rises rapidly with viral infection, then falls over days to weeks. Interleukin-8 correlates well with symptoms and can be reliably measured.129, 130, 131, 132, 133, 134, 135, 136, 137, 138 In our study, nasal wash samples are divided into aliquots, frozen at −80°C, and then analyzed for IL-8 in batches by enzyme-linked immunoassay, using previously described methods.139

Nasal mucus index 

Nasal wash samples are visually inspected and rated as: 0.5 = few gobs of mucus, 1.0 = one third of sample consists of mucus, 2.0 = two thirds of sample consists of mucus, 3.0 = all or nearly all of sample consists of mucus. This scale was developed by researchers at the Wisconsin State Laboratory of Hygiene and has not been extensively validated.

Temperature 

Body temperature is assessed by digital oral thermometer at enrollment and at day 3.

Timing the visit 

The study physician will start a stopwatch when he/she enters the exam room, and stop it when he/she exits. This will allow an assessment of the duration of both standard and enhanced clinical interactions, which could be useful for future cost-benefit analyses.

Procedures and Participant Flow 

We advertise widely in the Dane County community by using flyers, posters, mailings, newspaper ads, email messages, and presentations to community groups. Prospective participants call a listed phone number and are screened by a research specialist. Those who are eligible and willing come in for informed consent and enrollment. At enrollment, participants fill out intake forms, are trained on the outcomes instruments, and for the 75% allocated to pill treatment, receive bottles of pills and take the first dose. Randomization is accomplished using allocation assignment cards in consecutively numbered sealed envelopes. For those assigned to a clinician visit (66.7% of the sample), a second smaller envelope with a card indicating interaction type is opened by the clinician at the exam room door. Randomization envelopes were prepared by the University of Wisconsin, Madison Investigational Drug Service by using a balanced blocks-of-twelve design. For those assigned to pill treatment, the first dose occurs at enrollment. Usually, nasal wash and instrument scoring is completed prior to clinician visit. During enrollment, the participant arranges for daily phone contact and for a day 3 follow-up visit. On each subsequent day of illness, and for two days after resolution, the participant self-administers outcomes instruments at home. The exit interview is scheduled as soon as practical after the participant has indicated that he/she is not sick for two days in a row. The exit interview includes a final questionnaire, an inventory of potential side effects, and return of the pill bottles for pill-counting adherence assessment.

Nonprotocol treatments 

Participants are asked to refrain from using other treatments (eg, antibiotics, antivirals, decongestants, antihistamines, cough suppressants, expectorants, antiinflammatories, combination cold formulas, nasal steroids, anticholinergics). In the case of significant pain (subjective rating of five or more on a conventional zero to 10 scale), acetaminophen is allowed. If the participant feels the need for medical diagnosis or treatment, those consultations and treatments are noted and analyzed as secondary outcomes and/or confounding variables. Treatments containing zinc or vitamin C are also disallowed, as are herbal treatments including garlic, ginger, Andrographis, Astragalus, or Echinacea. Other herbal teas (chamomile, peppermint) and home remedies (chicken soup) and nasal saline are allowed. Participants are asked about all nonprotocol treatments during daily telephone calls and at the exit interview.

Blinding 

Allocation to blinded echinacea and blinded placebo is concealed from participants, clinicians, research specialists, and investigators through the use of identical-appearing coated tablets and randomization codes. Testing of concealment is done by asking blinded participants to guess whether they received echinacea. Allocation to no pills and unblinded echinacea is not concealed from participants or researchers. Those responsible for conducting and reporting laboratory tests and for entering and cleaning data are blinded from all group assignments. Initial stages of statistical analysis and primary hypothesis testing will be done before breaking the blind. The study is advertised as a randomized trial testing echinacea as cold treatment. Study materials and consent procedures do not purposefully deceive participants or conceal the existence, nature, and purpose of randomizing people into the standard and enhanced groups. Emphasis, however, is placed on the study’s aim of testing the effectiveness of echinacea and on the clinician’s purpose of verifying diagnosis. Procedures and interactions are designed to focus participants’ attention on the honest and accurate reporting of their symptoms and not on the possible effects of placebo or physician interaction.

Adverse effects 

Theory and published evidence do not predict any specific adverse effects from any of the interventions. Although allergic response has been reported for Echinacea,140, 141 with an increased incidence of rash noted in one trial,142 the overall evidence does not suggest allergic potential beyond that of many common foods, or any other specific adverse effects.43, 143, 144, 145 In our study, adverse effects are monitored by asking the participant during the daily interviews “Do you think that you have had any side effects from the pills you’ve taken?” If the participant answers yes, the research assistant then says, “Please tell me about that.” Responses are noted, categorized, and compared among echinacea and placebo groups. At the exit interview, we specifically ask about bad taste, diarrhea, headache, nausea, rash, and stomach upset. Adverse effect rates among interventions will be compared during the blinded phase of statistical analysis.

Human subjects 

A data and safety-monitoring plan was approved by the University of Wisconsin Institutional Review Board’s Human Subjects Committee. Participants sign written consent and are modestly compensated for their participation. An independent data and safety monitoring committee meets at least once yearly to monitor the trial’s progress and look for trends in reported side effects. This committee is composed of two physician researchers, a pharmacist, and a statistician, all of whom are faculty at the University of Wisconsin School of Medicine and Public Health, and none of whom have conflicts of interest with this study or this paper’s authors.

Statistical Analysis 

The balanced two-way treatment structure used in this trial allows estimation of main effects in both directions (pill-related, clinician-related), potential identification of interactions, and assessment of additional prespecified relationships and contrasts of interest.1 Potential effects on primary and secondary outcomes will be tested using analysis of variance, analysis of covariance, and various multivariate regression statistical methods.146, 147, 148, 149, 150, 151, 152 The general linear mixed model will form the final analysis structure for identifying and assessing effects on primary and secondary outcomes.153, 154, 155, 156, 157, 158, 159, 160, 161

Theory and/or empirical research suggest that the following covariates may influence main outcomes, and hence will be considered for inclusion in final models. Covariates and confounders were (1) time of year (seasonal or etiological effect), (2) severity and duration of symptoms at enrollment, (3) use of nonprotocol medications, (4) use of tobacco, (5) demographics (age, gender, ethnicity, income, education), (6) general health-related quality of life, (7) perceived stress at enrollment, (8) interpersonal support, (9) optimism, (10) positive/negative affectivity, and (11) expectation (belief in echinacea and placebo effects; prediction of length and severity of illness).

Sample size determinations are based on statistical power to detect clinically significant differences between groups.162, 163, 164, 165 Although there is no consensus regarding clinically significant effect size for the common cold, our own work suggests that a one-day reduction in duration and/or a 20% reduction in severity would be sufficient to justify treatment for many or most cold sufferers.166, 167, 168 This trial is powered to detect a 20% between-group difference in severity days for pill-related groups (n = 180) and a 15% difference among clinician-related groups (n = 240). Power calculations were based on β = .20 (80% power), α = .05, one-sided testing, and observed variability from our own previous research using the WURSS instrument. One-sided testing is justified by the published literature, virtually all of which suggests effects in positive directions.

Severity days are defined as the area under the time-severity curve, with the y-axis defined by WURSS-21 scores, and the x-axis defined as duration of illness, starting with first symptom and ending when the participant first indicates that he/she is no longer sick. Total severity days will be calculated for each subject. Distributions of severity days will be tested and possibly corrected for normality, then contrasted among groups by using statistical techniques mentioned above.

Evidence for treatment effects on biomarkers (IL-8, nasal neutrophils) and self-reported outcomes (quality of life, perceived stress, and positive and negative affectivity) will be analyzed using similar statistical techniques as for the primary outcome, but will be interpreted more cautiously due to multiple comparison considerations.

We plan to use false discovery rate methods169, 170, 171, 172 rather than Bonferroni or random field method corrections to account for multiple comparison considerations.

Central tendency and distribution of each variable will be assessed using standard graphical and tabular techniques. Bivariate relationships will be assessed using standard measures of association. Tests for linearity, independence, and normality will be conducted with consideration of transformation to better approximate statistical assumptions. Analysis of missing data will be conducted to assess the level and possible reasons, with consideration of imputation.173, 174 All decisions for possible transformation or imputation will be done prior to unblinding.

Intention-to-treat analysis preserves trial integrity by taking account of as much outcome data as possible for all individuals randomized. Several methods for dealing with missing values have been proposed.175, 176 An approach based on random effects pattern-mixture modeling is especially promising.173, 177 If type and degree of missing data warrants, we will code subjects based on missing-data patterns, then use these codes as covariates in final models. This will be done prior to unblinding.

This trial is designed in accordance with CONSORT reporting guidelines,178, 179, 180 is registered with the National Institutes of Health, CRISP, and Cochrane, and is in compliance with recommendations from the International Committee of Medical Journal Editors.181, 182

Back to Article Outline

Discussion 

Kaptchuk has described placebos as “the dark side of the randomised controlled trial,”183 the method of double blinding as “intentional ignorance,”184 and has asked whether the double blind RCT methodology is “gold standard or golden calf?”185 Such strong words highlight the controversial nature of placebos in modern healthcare theory and practice. On the one hand, placebo control is at the heart of our most rigorous method of determining whether medical interventions provide their claimed effects. On the other hand, placebos and placebo effects are questioned, minimized, derided, and eschewed.186, 187, 188 Clearly, as stated by Di Blasi and Reilly,189 placebos are “medical paradoxes [that] need disentangling.” Unfortunately, as we have tried to make clear in our background and theory article,19 the terms placebo and placebo effect have numerous meanings, many of which elude confirmation and measurement by using standard hypothetico-reductionist methodologies.

This trial was designed to shed light on placebo effects, both in terms of pill-taking and doctor-patient interaction. At this point, we see a few caveats regarding interpretation. First, despite meticulous planning and rigorous adherence to protocol, what we are testing may not be what we would truly want to assess. This is likely most true of the doctor-patient interaction where the somewhat artificial nature of the clinic visit may limit generalization. Although people participate in research for several reasons, the most common motivations may be curiosity, desire to contribute to science and public health, and compensation. In clinical practice, on the other hand, patients usually come seeking help for an illness condition, to ameliorate symptoms, enhance function, or simply find an explanation for what has been troubling them. In clinical context, the words and actions of doctors may have enhanced effects, especially if the doctor-patient relationship has evolved over time to include familiarity and trust. In our research setting, participants see the clinician-intervener (someone they have not yet met) for five to 10 minutes, generally with nontherapeutic motivations. Thus, this trial tests doctor-patient interaction in an acute setting but may not be generalizable to settings where patients know and trust their physicians.

Pill-related placebo effects may also differ in research and nonresearch settings. The PEP trial is designed to assess effects between real world choices (not taking a pill vs taking a pill that is known to have active ingredients) and research-setting possibilities (taking a pill that has a 50% chance of being placebo or active treatment. This is all done within a research setting, however, which may to some extent limit generalization. People entering research settings are not blank slates. They come with knowledge, attitudes, and experience that may influence response patterns. Someone with positive beliefs regarding echinacea, for example, may respond more favorably to being assigned to open-label echinacea (compared with blinded echinacea) than someone with less favorable views.

At the time of writing, we have little clue as to what the final results will be, due to blinding considerations. We have, however, learned a few things. For the clinician-interveners who normally practice medicine in patient-oriented enhanced fashion, we learned how difficult it is to emulate a standard interaction, consciously avoiding the warmth, reassurance, and positive talk that we employ in our clinical work. We more fully understand that clinical trials are artificial settings and feel even more strongly than before we began that the nature of the placebo effect is subtle, complex, and not well suited to elucidation through conventional methodology.

Our research raises a basic question about the healer’s mode of impact through either the dispensed product and/or through the nature of the encounter itself. Although clinical drug trials focus on the drug product as the primary vehicle for improved outcomes, with other factors characterized derogatorily as placebo effect, what if the converse is true? Might it be equally justified to imagine the healer’s interaction style to be a primary mode of healing, with the product’s impact secondary? Answers to the question of whether placebo or context effects represent powerful therapies or methodological bias32 may never be entirely satisfactory. Perhaps this is as it should be, as the study of placebo, perhaps more than any other field of medical science, takes us to the frontiers and borderlands of health, mind, and body, where linear cause-and-effect hypothesis testing seems clumsy and overly simplified and unlikely to provide the understanding that we seek. Nevertheless, the importance to medical research and practice can hardly be overstated, and it is this tremendous potential that calls us toward further exploration. We truly hope that the results of the research described here will add knowledge and enhance understanding of this vital field of inquiry, so as to contribute toward the advancement of human health.

Back to Article Outline

Acknowledgment 

We thank the hundreds of trial participants who graciously gave of their time, energy, and goodwill to support medical research during a time of illness.

Back to Article Outline

References 

  1. Friedman LM, Furberg CD, DeMets DL. Fundamentals of Clinical Trials. New York, NY: Springer-Verlag; 1998;
  2. Guyatt GH, Rennie D. Users’ Guides to the Literature: A Manual for Evidence-Based Clinical Practice. Chicago, Ill: AMA Press; 2002;
  3. Benson K, Hartz AJ. A comparison of observational studies and randomized, controlled trials. N Engl J Med. 2000;342:1878–1886
  4. Concato J, Shah N, Horwitz RI. Randomized, controlled trials, observational studies, and the hierarchy of research designs. N Engl J Med. 2000;342:1887–1892
  5. Kunz R, Oxman AD. The unpredictability paradox: review of empirical comparisons of randomised and non-randomised clinical trials. BMJ. 1998;317:1185–1190
  6. McKee M, Britton A, Black N, McPherson K, Sanderson C, Bain C. Methods in health services research (Interpreting the evidence: choosing between randomised and non-randomised studies). BMJ. 1999;319:312–315
  7. Ottenbacher K. Impact of random assignment on study outcome: an empirical examination. Control Clin Trials. 1992;13:50–61
  8. Pocock SJ, Elbourne DR. Randomized trials or observational tribulations?. N Engl J Med. 2000;342:1907–1909
  9. Basoglu M, Marks I, Livanou M, Swinson R. Double-blindness procedures, rater blindness, and ratings of outcome: observations from a controlled trial. Arch Gen Psychiatry. 1997;54:744–748
  10. Berlin JA, Rennie D. Measuring the quality of trials: the quality of quality scales. JAMA. 1999;282:1083–1085
  11. Caspi O, Millen C, Sechrest L. Integrity and research: introducing the concept of dual blindness (How blind are double-blind clinical trials in alternative medicine?). J Altern Complement Med. 2000;6:493–498
  12. Devereaux PJ, Manns BJ, Ghali WA, et al. Physician interpretations and textbook definitions of blinding terminology in randomized controlled trials. JAMA. 2001;285:2000–2003
  13. Jadad-Bechara AR, Moore RA, Carrol D. Assessing the quality of reports of randomized clinical trials: is blinding necessary?. Control Clin Trials. 1996;17:1–12
  14. Engel LW, Guess HA, Kleinman A, Kusek JW. The Science of the Placebo: Toward an Interdisciplinary Research Agenda. London, England: BMJ Books; 2002;
  15. Harrington A. The Placebo Effect: An Interdisciplinary Exploration. Cambridge, Mass: Harvard University Press; 1997;
  16. Moerman DE. Meaning, Medicine and the ‘Placebo Effect’. Cambridge, NY: Cambridge University Press; 2002;
  17. Kirsch I. How Expectancies Shape Experience. Washington, DC: American Psychological Association; 1999;
  18. Shapiro AK, Shapiro E. The Powerful Placebo: From Ancient Priest to Modern Physician. Baltimore, Md: Johns Hopkins University Press; 1997;
  19. Barrett B, Muller D, Rakel D, Rabago D, Marchand L, Scheder J. Placebo, meaning and health. Perspect Biol Med. 2006;49:178–198
  20. Grufferman S. Complexity and the Hawthorne effect in community trials. Epidemiology. 1999;10:209–210
  21. Hsueh Y. The Hawthorne experiments and the introduction of Jean Piaget in American industrial psychology, 1929-1932. Hist Psychol. 2002;5:163–189
  22. Wickstrom G, Bendix T. The “Hawthorne effect”–what did the original Hawthorne studies actually show?. Scand J Work Environ Health. 2000;26:363–367
  23. Mead N, Bower P. Patient-centredness: a conceptual framework and review of the empirical literature. Soc Sci Med. 2000;51:1087–1110
  24. Flocke SA, Miller WL, Crabtree BF. Relationships between physician practice style, patient satisfaction, and attributes of primary care. J Fam Pract. 2002;51:835–840
  25. Moerman D. Doctors and patients: the role of clinicians in the placebo effect. Adv Mind Body Med. 2003;19:14–22
  26. Roter DL, Hall JA. Studies of doctor-patient interaction. Ann Rev Public Health. 1989;10:163–180
  27. Stewart MA. Effective physician-patient communication and health outcomes: a review. Can Med Assoc J. 1995;152:1423–1433
  28. Barsky AJ, Saintfort R, Rogers M, Borus JF. Nonspecific medication side effects and the nocebo phenomenon. JAMA. 2002;287:622–627
  29. Bouchet C, Guillemin F, Briancon S. Nonspecific effects in longitudinal studies: impact on quality of life measures. J Clin Epidemiol. 1996;49:15–20
  30. Kirsch I. Specifying nonspecifics: psychological mechanisms of placebo effects. In:  Harrington A editors. The Placebo Effect: An Interdisciplinary Exploration. Cambridge Mass: Harvard University Press; 1997;p. 166–186
  31. Di Blasi Z, Harkness E, Ernst E, Georgiou A, Kleijnen J. Influence of context effects on health outcomes: a systematic review. Lancet. 2001;357:757–762
  32. Di Blasi Z, Kleijnen J. Context effects (Powerful therapies or methodological bias?). Eval Health Prof. 2003;26:166–179
  33. Thomas KB. General practice consultations: is there any point in being positive?. BMJ. 1987;294:1200–1203
  34. Barrett BP, Brown RL, Locken K, Maberry R, Bobula JA, D’Alessio D. Treatment of the common cold with unrefined echinacea: a randomized, double-blind, placebo-controlled trial. Ann Intern Med. 2002;137:939–946
  35. Barrett B, Locken K, Maberry R, et al. The Wisconsin Upper Respiratory Symptom Survey: development of an instrument to measure the common cold. J Fam Pract. 2002;51:265–273
  36. Barrett B, Brown R, Mundt M, et al. The Wisconsin Upper Respiratory Symptom Survey is responsive, reliable, and valid. J Clin Epidemiol. 2005;58:609–617
  37. Barrett B, Brown R, Voland R, Maberry R, Turner R. Relations among questionnaire and laboratory measures of rhinovirus infection. Eur Respir J. 2006;28:358–363
  38. Barrett B, Mundt MP, Brown RE. Assessing important difference and responsiveness in common cold. Qual Life Res. In press.
  39. Barrett B, Vohmann M, Calabrese C. Echinacea for upper respiratory infection: evidence-based clinical review. J Fam Pract. 1999;48:628–635
  40. Caruso TJ, Gwaltney JM. Treatment of the common cold with echinacea: a structured review. Clin Infect Dis. 2005;40:807–810
  41. Giles JT, Palat CT, Chien SH, Chang ZG, Kennedy DT. Evaluation of Echinacea for treatment of the common cold. Pharmacotherapy. 2000;20:690–697
  42. Kligler B. Echinacea. Am Fam Physician. 2003;67:77–80
  43. Linde K, Barrett B, Wolkart K, Bauer R, Melchart D. Echinacea for preventing and treating the common cold. Cochrane Database Syst Rev. 2006;(1):CD000530
  44. Melchart D, Linde K, Fischer P, Kaesmayr J. Echinacea for preventing and treating the common cold. Cochrane Library. 2000;(Issue 4):
  45. Barrett B, Locken K, Maberry R, et al. The Wisconsin Upper Respiratory Symptom Survey: development of an instrument to measure the common cold. J Fam Pract. 2002;51:265–273
  46. Jackson GG, Dowling HF, Spiesman IG, Boand AV. Transmission of the common cold to volunteers under controlled conditions. Arch Intern Med. 1958;101:267–278
  47. Jackson GG, Dowling HF, Anderson TO, Riff L, Saporta J, Turck M. Susceptibility and immunity to common upper respiratory viral infections-the common cold. Ann Intern Med. 1960;55:719–738
  48. Jackson GG, Dowling HF, Muldoon RL. Present concepts of the common cold. Am J Public Health. 1962;52:940–945
  49. Balint M. Is the therapeutic nature of the patient-physician relationship being undermined?. Arch Intern Med. 2000;160:2257–2259
  50. Bass MJ, Buck C, Turner L, Dickie G, Pratt G, Robinson HC. The physician’s actions and the outcome of illness in family practice. J Fam Pract. 1986;23:43–47
  51. Brody H. Placebo response, sustained partnership, and emotional resilience in practice. J Am Board Fam Pract. 1997;10:72–74
  52. Charon R. The patient-physician relationship (Narrative medicine: a model for empathy, reflection, profession and trust). JAMA. 2001;286:1897–1902
  53. Chewning B, Sleath B. Medication decision-making and management: a client-centered model. Soc Sci Med. 1996;42:389–398
  54. Egbert LD, Battit GE, Welch CE, Bartlett MK. Reduction of post-operative pain by encouragement and instruction of patients. NEJM. 1964;270:825–827
  55. Kaplan SH, Greenfield S, Ware JE. Assessing the effects of physician-patient interactions on the outcomes of chronic disease. Med Care. 1989;27:S110–S127
  56. Krupat E, Bell RA, Kravitz RL, Thom D, Azari R. When physicians and patients think alike: patient-centered beliefs and their impact on satisfaction and trust. J Fam Pract. 2001;50:1057–1062
  57. Latham CE. Is there data to support the concept that educated, empowered patients have better outcomes?. J Am Soc Nephrol. 1998;9(suppl 4):S141–S144
  58. Levenstein JH, McCracken EC, McWhinney IR, Stewart MA, Brown JB. The patient-centred clinical method (1. A model for the doctor-patient interaction in family medicine). Fam Pract. 1986;3:24–30
  59. Pappas S, Perlman A. Complementary and alternative medicine (The importance of doctor-patient communication). Med Clin North Am. 2002;86:1–10
  60. Reilly D. Enhancing human healing. BMJ. 2001;322:120–121
  61. Roter DL, Hall JA, Kern DE, Barker LR, Cole KA, Roca RP. Improving physicians’ interviewing skills and reducing patients’ emotional distress (A randomized clinical trial). Arch Intern Med. 1995;155:1877–1884
  62. Simpson M, Buckman R, Stewart M, et al. Doctor-patient communication: the Toronto consensus statement. BMJ. 1991;303:1385–1387
  63. Starfield B, Wray C, Hess K, Gross R, Birk PS, D’Lugoff BC. The influence of patient-practitioner agreement on outcome of care. Am J Public Health. 1981;71:127–131
  64. Stewart M, Brown JB, Boon H, Galajda J, Meredith L, Sangster M. Evidence on patient-doctor communication. Cancer Prev Control. 1999;3:25–30
  65. Suchman AL, Markakis K, Beckman HB, Frankel R. A model of empathic communication in the medical interview. JAMA. 1997;277:678–682
  66. Thomas KB. The consultation and the therapeutic illusion. BMJ. 1978;1:1327–1328
  67. Roter DL, Hall JA, Merisca R, Nordstrom B, Cretin D, Svarstad B. Effectiveness of interventions to improve patient compliance: a meta-analysis. Med Care. 1998;36:1138–1161
  68. Roter DL, Hall JA. Doctors Talking with Patients: Patients Talking with Doctors: Improving Communication in Medical Visits. Westport Conn: Auburn House; 1992;
  69. Flood AB, Lorence DP, Ding J, McPherson K, Black NA. The role of expectations in patients’ reports of post-operative outcomes and improvement following therapy. Med Care. 1993;31:1043–1056
  70. Little P, Everitt H, Williamson I, et al. Observational study of effect of patient centredness and positive approach on outcomes of general practice consultations. BMJ. 2001;323:908–911
  71. Mondloch MV, Cole DC, Frank JW. Does how you do depend on how you think you’ll do? (A systematic review of the evidence for a relation between patients’ recovery expectations and health outcomes). Can Med Assoc J. 2001;165:174–179
  72. Petrie KJ, Weinman J, Sharpe N, Buckley J. Role of patients’ view of their illness in predicting return to work and functioning after myocardial infarcation: longitudinal study. BMJ. 1996;312:1191–1194
  73. Ryff CD. Positive mental health. In:  Blechman EA,  Brownell KD editor. Behavioral Medicine and Women: A Comprehensive Handbook. New York, NY: Guilford Publications, Inc; 1998;p. 183–188
  74. Savage R, Armstrong D. Effect of a general practitioner’s consulting style on patients’ satisfaction: a controlled study. BMJ. 1990;301:968–970
  75. Roter DL, Hall JA. Patient-provider communication. In:  Glanz K,  Lewis FM,  Rimer BK editor. Health Behavior and Health Education: Theory, Research and Practice. San Francisco, Calif: Jossey-Bass, A Wiley Company; 1997;p. 206–226
  76. Evans RG, Barer ML, Marmot TR. Why are Some People Healthy and Others Not? (The Determinants of Health in Populations). New York, NY: Aldine de Gruyter; 1994;
  77. Marmot MG. Social differentials in health within and between populations. Daedalus: J Am Acad Arts Sci. 1994;123:197–216
  78. McKnight JL. Health and empowerment. Can J Public Health. 1985;(suppl 1):37–38
  79. Singer B, Ryff CD. Hierarchies of life histories and associated health risk. Ann N Y Acad Sci. 1999;896:95–115
  80. Guadagnoli E, Ward P. Patient participation in decision-making. Soc Sci Med. 1998;47:329–339
  81. Suchman AL, Matthews DA. What makes the patient-doctor relationship therapeutic? (Exploring the connexional dimension of medical care). Ann Intern Med. 1988;108:125–130
  82. Matthews DA, Suchman AL, Branch WT. Making “connexions”: enhancing the therapeutic potential of patient-clinician relationships. Ann Intern Med. 1993;118:973–977
  83. Goldberg PE. The physician-patient relationship: three psychodynamic concepts that can be applied to primary care. Arch Fam Med. 2000;9:1164–1168
  84. Mercer SW, Reynolds WJ. Empathy and quality of care. Br J Gen Pract. 2002;52(suppl):S9–S12
  85. Mercer SW, McConnachie A, Maxwell M, Heaney D, Watt GCM. Relevance and practical use of the Consultation and Relational Empathy (CARE) Measure in general practice. Fam Pract. 2005;22:328–334
  86. Mercer SW, Maxwell M, Heaney D, Watt GCM. The consultation and relational empathy (CARE) measure: development and preliminary validation and reliability of an empathy-based consultation process measure. Fam Pract. 2004;21:699–705
  87. Hall JA, Horgan TG, Stein TS, Roter DL. Liking in the physician–patient relationship. Patient Educ Couns. 2002;48:69–77
  88. Badia LX, Herdman M, Schiaffino A. Determining correspondence between scores on the EQ-5D “thermometer” and a 5-point categorical rating scale. Med Care. 1999;37:671–677
  89. Coyle D, Wells G, Graham I, Lee KM, Peterson JE, Papadimitropoulos E. The impact of risk on preference values: implications for evaluations of postmenopausal osteoporosis therapy. Value Health. 2001;4:385–391
  90. Feeny D, Furlong W, Torrance GW, et al. Multiattribute and single-attribute utility functions for the health utilities index mark 3 system. Med Care. 2002;40:113–128
  91. Brazier JE, Harper R, Jones NMB, et al. Validating the SF-36 health survey questionnaire: new outcome measure for primary care. BMJ. 1992;305:160–164
  92. Jenkinson C, Wright L, Coulter A. Criterion validity and reliability of the SF-36 in a population sample. Qual Life Res. 1994;3:7–12
  93. Ware JE, Sherbourne CD. The MOS 36-item short-form health survey (SF-36): I (Conceptual framework and item selection). Med Care. 1992;30:473–483
  94. McHorney CA, Ware JE, Raczek AE. The MOS 36-item short-form health survey (SF-36): II (Psychometric and clinical tests of validity in measuring physical and mental health constructs). Med Care. 1998;31:247–263
  95. McHorney CA, Ware JE, Lu JFR, Sherbourne CD. The MOS 36-Item short-form health survery (SF-36): III (Tests of data quality, scaling assumptions, and reliability across diverse patient groups). Med Care. 1994;32:40–66
  96. Ware JE, Kosinski M, Dewey JE, Gandek B. How to Score and Interpret Single-Item Health Status Measures: A Manual for Users of the SF-8 Health Survey. Lincoln, RI: QualityMetric; 2001;
  97. Cobb JM, Steptoe A. Psychosocial stress and susceptibility to upper respiratory tract illness in an adult population sample. Psychosom Med. 1996;58:404–412
  98. Cohen S, Tyrell DAJ, Smith AP. Psychological stress and susceptibility to the common cold. NEJM. 1991;325:606–612
  99. Stone AA, Bovbjerg DH, Neale JM, et al. Development of common cold symptoms following experimental rhinovirus infection is related to prior stressful life events. Behav Med. 1992;18:115–120
  100. Takkouche B, Regueira C, Gestal-Otero JJ. A cohort study of stress and the common cold. Epidemiology. 2001;12:345–349
  101. Labbate LA, Fava M, Oleshansky M, Zoltec J, Littman A, Harig P. Physical fitness and perceived stress (Relationships with coronary artery disease risk factors). Psychosomatics. 1995;36:555–560
  102. Malarkey WB, Pearl DK, Demers LM, Kiecolt-Glaser JK, Glaser R. Influence of academic stress and season on 24-hour mean concentrations of ACTH, cortisol, and beta-endorphin. Psychoneuroendocrinology. 1995;20:499–508
  103. van Eck M, Nicolson NA. Perceived stress and salivary cortisol in daily life. Ann Behav Med. 1994;16:221–227
  104. van Eck M, Berkhof H, Nicolson N, Sulon J. The effects of perceived stress, traits, mood states, and stressful daily events on salivary cortisol. Psychosom Med. 1996;58:447–458
  105. Cohen S, Underwood LG, Gottlieb BH. Social Support Measurement and Intervention. Oxford, UK: Oxford University Press; 2000;
  106. Goodwin P, Leszcz M, Ennis M, et al. The effect of group psychosocial support on survival in metastatic breast cancer. NEJM. 2001;345:1719–1726
  107. Ryff CD, Singer B. Interpersonal flourishing: a positive health agenda for the new millenium. Pers Soc Psychol Rev. 2000;4:30–44
  108. Cohen S, Doyle WJ, Skoner DP, Rabin BS, Gwaltney JM. Social ties and susceptibility to the common cold. JAMA. 1997;277:1940–1944
  109. Ryff CD, Singer B. Psychological well-being: meaning, measurement, and implications for psychotherapy research. Psychother Psychosom. 1996;65:14–23
  110. Desharnais R, Godin G, Jobin J, Valois P, Ross A. Optimism and health-relevant cognitions after a myocardial infarction. Psychol Rep. 1990;67:t-5
  111. Ryff CD. Psychological well-being in adult life. Curr Dir Psychol Sci. 1995;4:99–104
  112. Schmutte PS, Ryff CD. Personality and well-being: reexamining methods and meanings. J Pers Soc Psychol. 1997;73:549–559
  113. Scheier MF, Carver CS. Optimism, coping, and health: assessment and implications of generalized outcome expectancies. Health Psychol. 1985;4:219–247
  114. Scheier MF, Carver CS. Effects of optimism on psychological and physical well-being: theoretical overview and empirical update. Cognit Ther Res. 1992;16:201–228
  115. Cohen S, Tyrrell DAJ, Smith AP. Negative life events, perceived stress, negative affect, and susceptibility to the common cold. J Pers Soc Psychol. 1993;64:131–140
  116. Cohen S, Doyle WJ, Skoner DP, Fireman P, Gwaltney JM, Newsom JT. State and trait negative affect as predictors of objective and subjective symptoms of respiratory viral infections. J Pers Soc Psychol. 1995;68:159–169
  117. Cohen S, Alper CM, Doyle WJ, Treanor JJ, Turner RB. Positive emotional style predicts resistance to illness after experimental exposure to rhinovirus or influenza a virus. Psychosom Med. 2006;68:809–815
  118. Albrecht RR, Ewing SJ. Standardizing the administration of the Profile of Mood States (POMS): development of alternative word lists. J Pers Assess. 1989;53:31–39
  119. Jacobs KW, Boze MM. Correlations among scales of the Beck Depression Inventory and the profile of mood states. Psychol Rep. 1993;73:431–434
  120. Little K, Penman E. Measuring subacute mood changes using the profile of mood states and visual analogue scales. Psychopathology. 1989;22:42–49
  121. Nyenhuis DL, Yamamoto C, Luchetta T, Terrien A, Parmentier A. Adult and geriatric normative data and validation of the profile of mood states. J Clin Psychol. 1999;55:79–86
  122. Pollock V, Cho DW, Reker D, Volavka J. Profile of Mood States: the factors and their physiological correlates. J Nerv Ment Dis. 1979;167:612–614
  123. Shacham S. A shortened version of the Profile of Mood States. J Pers Assess. 1983;47:305–306
  124. Abbas AK, Lichtman AH, Pober JS. Cellular and Molecular Immunology. Philadelphia, Pa: W.B. Saunders; 2000;
  125. Janeway CA, Travers P, Walport M, Capra JD. Immunobiology: the Immune System in Health and Disease. New York, NY: Current Biology Publications; 1999;
  126. Naclerio RM, Proud D, Lichtenstein LM, et al. Kinins are generated during experimental rhinovirus colds. J Infect Dis. 1988;157:133–142
  127. Teran LM, Johnston SL, Schroder JM, Church MK, Holgate ST. Role of nasal interleukin-8 in neutrophil recruitment and activation in children with virus induced asthma. Am J Respir Crit Care Med. 1997;155:1362–1366
  128. Turner RB. The treatment of rhinovirus infections: progress and potential. Antiviral Res. 2001;49:1–14
  129. Becker S, Koren HS, Henke DC. Interleukin-8 expression in normal nasal epithelium and its modulation by infection with respiratory syncytial virus and cytokines tumor necrosis factor, interleukin-1, and interleukin-6. Am J Respir Cell Mol Biol. 1993;8:20–27
  130. Kaul P, Biagioli MC, Singh I, Turner RB. Rhinovirus-induced oxidative stress and interleukin-8 elaboration involves p47-phox but is independent of attachment to intercellular adhesion molecule-1 and viral replication. J Infect Dis. 2000;181:1885–1890
  131. Choi AMK, Jacoby DB. Influenza virus A infection induces interleukin-8 gene expression in human airway epithelial cells. FEBS. 1992;309:327–329
  132. Garofalo R, Patel JA, Sim C, Schmalstieg FC, Goldman AS. Production of cytokines by virus-infected human respiratory epithelial cells. J Allergy Clin Immunol. 1993;91:177
  133. Gern JE, Vrtis R, Kelly EAB, Dick EC, Busse WW. Rhinovirus produces nonspecific activation of lymphocytes through a monocyte-dependent mechanism. J Immunol. 1996;157:1605–1612
  134. Johnston SL, Papi A, Monick MM, Hunninghake GW. Rhinoviruses induce interleukin-8 mRNA and protein production in human monocytes. J Infect Dis. 1997;175:323
  135. Noah TL, Henderson FW, Wortman IA, et al. Nasal cytokine production in viral acute upper respiratory infection of childhood. J Infect Dis. 1995;171:584–592
  136. Parry DE, Busse WW, Sukow KA, Dick CR, Swenson C, Gern JE. Rhinovirus-induced PBMC responses and outcome of experimental infection in allergic subjects. J Allergy Clin Immunol. 2000;105:692–698
  137. Roseler S, Holtappels G, Wagenmann M, Bachert C. Elevated levels of interleukins IL-1 beta, IL-6 and IL-8 in naturally acquired viral rhinitis. Eur Arch Otorhinolaryngol. 1995;1:S61–S63
  138. Zhu Z, Tang W, Gwaltney JM, Wu Y, Elias JA. Rhinovirus stimulation of interleukin-8 in vivo and in vitro: role of NF-kB. Am J Physiol. 1997;273:814–824
  139. Schroth MK, Grimm E, Frindt P, et al. Rhinovirus replication causes RANTES production in primary bronchial epithelial cells. Am J Respir Cell Mol Biol. 1999;20:1220–1228
  140. Mullins RJ, Heddle R. Adverse reactions associated with echinacea: the Australian experience. Ann Allergy Asthma Immunol. 2002;88:42–51
  141. Soon SL, Crawford RI. Recurrent erythema nodosum associated with Echinacea herbal therapy. J Am Acad Dermatol. 2001;44:298–299
  142. Taylor JA, Weber W, Standish L, et al. Efficacy and safety of echinacea in treating upper respiratory tract infections in children: a randomized controlled trial. JAMA. 2003;290:2824–2830
  143. Barrett B. Echinacea: a safety review. HerbalGram: J Am Botan. 2003;57:36–39
  144. Huntley AL, Thompson CJ, Ernst E. The safety of herbal medicinal products derived from echinacea species: a systematic review. Drug Safe. 2005;28:387–400
  145. Perri D, Dugoua JJ, Mills E, Koren G. Safety and efficacy of echinacea (E. angustafolia, e. purpurea and e. pallida) during pregnancy and lactation. Can J Clin Pharmacol. 2006;13:e262–e267
  146. Bentler PM. Multivariate analysis with latent variables: causual modeling. In:  Rosenzweig MR,  Porter LW editor. Annual Review of Psychology. Stanford, Conn: Annual Reviews, Inc; 1980;p. 419–456
  147. Browne MW. Covariance structures. In:  Hawkins DM editors. Topics in Applied Multivariate Analysis. Cambridge, Mass: Cambridge University Press; 1982;p. 72–141
  148. Dillon WR, Goldstein M. Multivariate Analysis: Methods and Applications. New York, NY: John Wiley & Sons; 1984;
  149. Huitema BE. The Analysis of Covariance and Alternatives. New York, NY: John Wiley & Sons; 1980;
  150. Joreskog KA. Analysis of covariance structures. Scand J Stat. 1981;8:65–92
  151. Lewontin R. Annotation: the analysis of variance and the analysis of causes. Am J Hum Genetics. 1974;26:400–411
  152. Snedecor GW, Cochran WG. Statistical Methods. Ames, IA: Iowa State University Press; 1980;
  153. Albert JM, DeMets DL. On a model-based approach to estimating efficacy in clinical trials. Stat Med. 1994;13:2323–2335
  154. Brown RL. Assessing specific mediational effects in complex theoretical models. Struct Eq Model. 1997;4:142–156
  155. Bryk AS, Raudenbush SW. Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, Calif: Sage; 1992;
  156. Davis CS. Semi-parametric and non-parametric methods for the analysis of repeated measurements with applications to clinical trials. Stat Med. 1991;1:1959–1980
  157. Goldstein H. Multilevel Statistical Models. New York, NY: Halsted; 1995;
  158. Hedeker D, Gibbons RD. A random-effects ordinal regression model for multilevel analysis. Biometrics. 1994;50:933–944
  159. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika. 1986;73:13–22
  160. Littell RC, Milliken GA, Wolfinger RD. SAS System for Mixed Models. Cary, NC: SAS Institute, Inc; 1996;
  161. Stucki G, Daltroy L, Katz JN, Johannesson M, Liang MH. Interpretation of change scores in ordinal clinical scales and health status measures: the whole may not equal the sum of the parts. J Clin Epidemiol. 1995;49:711–717
  162. Friedman LM, Furberg CD, DeMets DL. Sample size. In: Fundamentals of Clinical Trials. New York, NY: Springer-Verlag; 1998;p. 94–129
  163. Gong J, Pinheiro JC, DeMets DL. Estimating significance level and power comparisons for testing multiple endpoints in clinical trials. Control Clin Trials. 2000;21:313–329
  164. Jaeschke R, Singer J, Guyatt GH. Measurement of health status: ascertaining the minimal clinically important difference. Control Clin Trials. 1989;10:407–415
  165. Guyatt GH, Bombardier C, Tugwell P. Measuring disease-specific quality of life in clinical trials. Can Med Assoc J. 1986;134:889–894
  166. Barrett B, Brown R, Mundt M, et al. Using benefit harm tradeoffs to estimate sufficiently important difference: the case of the common cold. Med Decis Making. 2005;25:47–55
  167. Barrett B, Brown D, Mundt M, Brown R. Sufficiently important difference: expanding the framework of clinical significance. Med Decis Making. 2005;25:250–261
  168. Barrett B, Harahan B, Brown D, Zhang Z, Brown R. Sufficiently important difference for common cold: severity reduction. Ann Fam Med. 2007;5:216–223
  169. Benjamini Y, Hochberg Y. On the adaptive control of the false discovery rate in multiple testing with independent statistics. J Educ Behav Stat. 2000;25:60–83
  170. Benjamini Y, Yekutieli D. Quantitative trait Loci analysis using the false discovery rate. Genetics. 2005;171:783–790
  171. Benjamini Y, Hochberg Y. Controlling the false discovery rate–a practical and powerful approach to multiple testing. J Roy Stat Soc. 1995;B57:289–300
  172. Hochberg Y, Benjamini Y. More powerful procedures for multiple significance testing. Stat Med. 1990;9:811–818
  173. Hedeker D, Gibbons RD. Application of random-effects pattern-mixture models for missing data in longitudinal studies. Psychol Methods. 1997;2:64–78
  174. Schafer JL. Analysis of Incomplete Multivariate Data. London, England: Chapman & Hall; 1997;
  175. Unnebrink K, Windeler J. Intention-to-treat methods for dealing with missing values in clinical trials of progressively deteriorating diseases. Stat Med. 2001;20:3931–3946
  176. Touloumi G, Babiker AG, Pocok SJ, Darbyshire JH. Impact of missing data due to drop-outs on estimators for rates of change in longitudinal studies A simulation study. Stat Med. 2001;20:3715–3728
  177. Little RJA. Modeling the drop-out mechanism in repeated-measures studies. J Am Stat Assoc. 1995;90:1112–1121
  178. Altman DG. Better reporting of randomised controlled trials: the CONSORT statement. BMJ. 1996;313:570–571
  179. Moher D, Schulz KF, Altman D. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. JAMA. 2001;285:1987–1991
  180. Ioannidis JP, Evans SJ, Gotzsche PC, et al. Better reporting of harms in randomized trials: an extension of the CONSORT statement. Ann Intern Med. 2004;141:781–788
  181. De Angelis CD, Drazen JM, Frizelle FA, et al. Is this clinical trial fully registered?–a statement from the International Committee of Medical Journal Editors. N Engl J Med. 2005;352:2436–2438
  182. Dickersin K, Rennie D. Registering clinical trials. JAMA. 2003;290:516–523
  183. Kaptchuk TJ. Powerful placebo: the dark side of the randomised controlled trial. Lancet. 1998;351:1722–1725
  184. Kaptchuk TJ. Intentional ignorance: a history of blind assessment and placebo controls in medicine. Bull Hist Med. 1998;72:389–433
  185. Kaptchuk TJ. The double-blind, randomized, placebo-controlled trial: gold standard or golden calf?. J Clin Epidemiol. 2001;54:541–549
  186. Kienle GS, Kiene H. The powerful placebo effect: fact or fiction?. J Clin Epidemiol. 1997;50:1311–1318
  187. Hrobjartsson A, Gotzsche PC. Is the placebo powerless? (An analysis of clinical trials comparing placebo with no treatment). N Engl J Med. 2001;344:1594–1602
  188. Walach H. The efficacy paradox in randomized controlled trials of CAM and elsewhere: beware of the placebo trap. J Altern Complement Med. 2001;7:213–218
  189. Di Blasi Z, Reilly D. Placebos in medicine: medical paradoxes need disentangling. BMJ. 2005;330:45

 This trial was supported by the National Center for Complementary and Alternative Medicine at the National Institutes of Health, grant NIH NCCAM 1-R01-AT-1428-01 (BB, JB, KB, RB, BC, DLD, LM, D. Rakel, D. Rabago, CDR). The University of Wisconsin School of Medicine and Public Health and the University of Wisconsin Department of Family Medicine have also contributed. Dr Barrett was supported by a K-23 career development grant from National Institutes of Health National Center for Complementary and Alternative Medicine a career development grant from the Robert Wood Johnson Foundation Generalist Physician Faculty Scholars Program.

 MediHerb supplied the echinacea and placebo tablets and conducted phytochemical analyses but did not contribute financially in any other way. MediHerb has no contractual right for involvement in data analysis, conclusions, or manuscript preparation.

PII: S1550-8307(07)00272-8

doi:10.1016/j.explore.2007.08.001

Explore: The Journal of Science and Healing
Volume 3, Issue 6 , Pages 561-572, November 2007

Article Tools