An Internet-Based Randomized, Placebo-Controlled Trial of Kava and Valerian for Anxiety and Insomnia : Medicine

Journal Logo


An Internet-Based Randomized, Placebo-Controlled Trial of Kava and Valerian for Anxiety and Insomnia

Jacobs, Bradly P. MD, MPH; Bent, Stephen MD; Tice, Jeffrey A. MD; Blackwell, Terri MA; Cummings, Steven R. MD, FACP

Author Information
Medicine 84(4):p 197-207, July 2005. | DOI: 10.1097/
  • Free


The herbal extracts kava and valerian are the leading dietary supplements used in the self-management of anxiety and insomnia, respectively. There is limited evidence to support their effectiveness for these common symptoms. The Internet has been used to a limited extent for research, but it is not known whether randomized controlled trials can be conducted entirely using Internet technology.

We performed a randomized, double-blind, placebo-controlled trial using a novel Internet-based design to determine if kava is effective for reducing anxiety and if valerian is effective for improving sleep quality. E-mail recruitment letters and banner advertisements on websites were used to recruit a large pool of interested participants (1551) from 45 states over an 8-week period. Participants were first asked to read study information, complete an online informed consent process, and undergo electronic identity verification. In order to be eligible for the study, participants were required to have 1) anxiety as documented by scores of at least 0.5 standard deviations above the mean on the State-Trait Anxiety Inventory State subtest (STAI-State) on 2 separate occasions, and 2) insomnia, defined as a "problem getting to sleep or staying asleep over the past 2 weeks." We randomly assigned 391 eligible participants to 1 of the following 3 groups, and mailed 28 days' supply: kava with valerian placebo (n = 121), valerian with kava placebo (n = 135), or double placebo (n = 135). The primary outcome measures were changes from baseline in anxiety (STAI-State questionnaire) and insomnia (Insomnia Severity Index [ISI]) compared with placebo.

Participants receiving placebo had a 14.4 point decrease in anxiety symptoms on the STAI-State score and an 8.3 point decrease in insomnia symptoms on the ISI. Those receiving kava had similar reductions in STAI-State score (2.7 point greater reduction in placebo compared with kava; 95% confidence interval [CI], −0.8 to +6.2). Those receiving valerian and placebo had similar improvements in sleep (0.4 point greater reduction in the placebo than the valerian group; 95% CI, −1.3 to +2.1). Results were similar when limited to the 83% of participants who adhered to study compounds for all 4 weeks.

Neither kava nor valerian relieved anxiety or insomnia more than placebo. This trial demonstrates the feasibility of conducting randomized, blinded trials entirely via the Internet.

Abbreviations: ISI = Insomnia Severity Index, STAI = State-Trait Anxiety Inventory.


Anxiety and insomnia are 2 of the most common symptoms among adults in the United States. In a recent national survey, 17% of adults met DSM III-R criteria for an anxiety disorder within the previous year22; milder forms of anxiety are even more common38. Anxiety is associated with medical illness39, reduced mental and physical function14, and economic costs of over $46 billion annually in the United States35. Insomnia affects approximately 35% of adults worldwide1,40 and is more common in the elderly and those with chronic medical conditions15,16,30,43. Insomnia is associated with decreased quality of life, poorer health, higher use of general medical services, decreased work productivity, and absenteeism2,25. These 2 conditions often coexist: up to 42% of patients with anxiety also suffer from insomnia30.

Two herbal remedies, kava (Piper methysticum) and valerian (Valerian officinalis) are widely available over the counter and are the most common dietary supplements used in the self-management of anxiety and insomnia, respectively37. Both herbs rank among the top 10 herbs sold in the United States and Europe; in the United States annual sales from both herbs are approximately $28 million8,36. The German Commission E (committee of the German Federal Institute for Drugs and Medical Devices) approved kava for use in conditions of "nervous anxiety," "stress," and "restlessness"; and valerian for "restlessness" and "sleeping disorders based on nervous conditions"7. Authors of a recent systematic review34 of 7 randomized controlled trials of kava6,23,28,44,48,50,51 concluded that the herb is superior to placebo for treating anxiety, but noted significant methodologic problems with the included studies. Authors of a systematic review46 of 9 randomized controlled trials of valerian, published in 8 papers,3,10,21,24,26,27,41,49 found contradictory findings and inconclusive evidence regarding efficacy for treating insomnia.

Given the widespread use of these herbal compounds in Europe and the United States and the methodologic limitations of previous research, we conducted a randomized, placebo-controlled, double-blind clinical trial to assess the efficacy of kava and valerian compared with placebo for the treatment of anxiety and insomnia symptoms using a novel, Internet-based design. Although the Internet has been used in clinical research4,5,47, to our knowledge this is the first study to test the feasibility of conducting a randomized blinded trial entirely over the Internet. This approach allowed us to study the effects of these herbal preparations on the self-management of anxiety and insomnia in the home setting.


This randomized, 3-arm, placebo-controlled, double-blind clinical trial was conducted entirely over the Internet. The University of California-San Francisco and the Western Institutional Review Board reviewed and approved the study protocol.

Study Participants

Participants were recruited through e-mail messages sent to subscribers of Alternative Medicine magazine who had asked to be notified about information pertaining to anxiety or insomnia. The trial was also advertised using banners placed on 2 websites that specialize in women's health.

Potential participants were directed to the study website and viewed information about kava, valerian, and the study procedures. People interested in proceeding further established an electronic signature and gave consent to provide their name, e-mail address, mailing address, and telephone number. This information was automatically confirmed using electronic services to verify the accuracy of name and address information (Experian Inc., Orange, CA). Potential participants then read the informed consent and were required to answer correctly 6 questions about the study procedures and potential risks and benefits in order to participate in the study. Consent was certified by the participant's electronic signature.

Potential participants completed an online questionnaire to determine eligibility. Adults who, on 2 separate occasions, reported having a "problem getting to sleep or staying asleep over the past 2 weeks" and scored at least 40 points (at least 0.5 standard deviation greater than the mean for the normal population) on the State subtest of the State-Trait Anxiety Inventory (STAI-State)45 were eligible for the study. Participants were required to be at least 21 years old, reside in the United States, and have access to e-mail and the World Wide Web.

Participants were queried for depressive symptoms using a valid and reliable 2-item instrument52, and caffeine consumption was assessed by asking the number of cups of coffee and tea and cans of soda consumed per day9.

Participants were excluded from the study if they reported using sedative-hypnotics or benzodiazepines, drinking more than 2 alcohol-containing drinks per day, having a history of liver disease, being pregnant, or breastfeeding.

Potential and actual participants were encouraged to contact study investigators using e-mail or the telephone to clarify issues pertaining to the study. Enrolled participants were asked to refrain from using kava and valerian during the study period. Courtesy letters were provided to all participants informing their primary care physician of the participant's involvement in the study and providing information on the study compounds and study procedures.

Randomization and Blinding

Eligible participants were randomly assigned to 1 of 3 treatment groups via a computerized random number generator. A central pharmacy packaged and distributed the study compounds to participants based on the randomization codes.

The participants and study investigators were blinded to treatment allocation. The study herbs and placebos were matched for color, texture, size, shape, and taste. Valerian is well known for its distinctive odor. To match the odor between valerian and valerian-placebo, capsules were stored together separated by a mesh lining for 2 weeks. To assess adequacy of treatment allocation concealment, participants were asked at the end of the trial if they thought they had received active treatment or placebo.


Study compounds and instructions were sent by overnight mail. Study investigators contacted participants by telephone or e-mail to verify receipt and to clarify questions. All study participants received 2 sets of capsules. Treatment groups received a 28-day course according to the following protocol:

  1. Kava group: 1 kava softgel capsule (each containing 100 mg of total kavalactones; 30% total kavalactones in extract) 3 times daily, and 2 placebo-valerian softgel capsules 1 hour before bedtime.
  2. Valerian group: 2 valerian softgel capsules (each containing 3.2 mg of valerenic acids; 1% valerenic acid in extract) 1 hour before bedtime, and 1 placebo-kava softgel capsule 3 times daily.
  3. Placebo group: these participants received both placebo-kava and placebo-valerian softgel capsules with the same dosing.

The excipient (excluding the extract) was soybean oil in the kava and placebo-kava capsules and fractionated coconut oil in the valerian and placebo-valerian capsules. Placebo capsules contained only the excipient with no active herbal extract. Study herbs and placebos were provided by PureWorld Botanicals (South Hackensack, NJ) and encapsulation was performed by RP Scherer Inc. (Basking Ridge, NJ). Kava and valerian capsules were each produced from 1 lot of each herb. Laboratory testing using high-performance liquid chromatography assay methodology19 of a random sample of softgel capsules found 100.2 mg total kavalactones in the active kava capsules and 3.2 mg valerenic acids in the active valerian capsules.

Outcome Measures

Data were collected at baseline and after 2 weeks and 4 weeks. The primary outcome measure in the assessment of kava was the change in anxiety symptoms after 4 weeks compared with placebo. This was assessed by the STAI-State subtest, a validated and reliable 20-item measure of anxiety symptoms45. The primary outcome in the assessment of valerian was the change in insomnia symptoms after 4 weeks compared with placebo. This was assessed by the Insomnia Severity Index (ISI), a 7-item instrument shown to be valid and reliable for measuring perceived insomnia severity31.

Secondary outcomes included change in insomnia symptoms using the ISI for kava and change in anxiety symptoms using the STAI-State subtest for valerian. Additional outcomes for valerian included the change in the frequency of nocturnal awakenings and change in sleep latency after 4 weeks compared with placebo.

To assess these outcomes, participants received automatic reminders by e-mail at 2 and 4 weeks of follow-up. The e-mail included an embedded universal resource locator (URL) to a secure website that contained follow-up instruments to assess anxiety and sleep. Adherence was assessed by self-report and pill count of bottles that were returned at the conclusion of the trial by prepaid overnight courier envelope. Participants did not receive monetary payment for their participation in the trial, but those who completed the trial were offered a free 1-month supply of their choice of kava or valerian.

Safety Monitoring

At 2- and 4-week follow-up, participants were required to complete an adverse event symptom inventory and were asked to report any hospitalizations, emergency room visits, or new symptoms or illnesses that required evaluation by a physician. Participants were also instructed to report this information throughout the study by contacting the on-call physician by e-mail or telephone. Adjudication of adverse events and safety data was performed by study physicians who were blinded to participant treatment allocation.

Sample Size and Statistical Analyses

Assuming a 23% decrease in the mean score on the STAI-State subtest within the placebo group, a 2-sided alpha of 0.05 and power of 80%, a sample size of 104 would have been needed for each group to detect a 10% greater reduction in the change in the STAI-State subtest between treatment and placebo group. Assuming a 33% decrease in the mean score on the ISI within the placebo group, a 2-sided alpha of 0.05 and power of 80%, a sample size of 104 would have been needed for each group to detect a 15% greater reduction in the change in the ISI between treatment and placebo group. Therefore, based on a 20% drop-out rate, we estimated that we would need to enroll at least 130 participants per treatment arm (390 participants for all 3 groups).

Differences in baseline characteristics between the 3 treatment arms were compared using chi-squared tests for categorical data and ANOVA for continuous data. The primary and secondary hypotheses were tested using a paired t-test for the difference in the change score between groups on the STAI-State subtest and on the ISI. The change score is defined by the difference between baseline and 4 week scores. For all significant differences (p < .05) in baseline characteristics between groups, a general linear model was used to adjust the final score. The primary outcome was the adjusted difference for both the STAI-State subtest and the ISI scores. The primary analyses compared the change in each herbal group to the change in the double placebo group.

The primary analyses included all participants who were randomized and provided follow-up information at 2 or 4 weeks regardless of the use of study compound ("intention to treat"). Subgroup analyses were also performed among participants who took at least 80% of their study compound during the study period.

To test the reliability and internal validity of participants' reports of symptoms, we analyzed the correlation between symptoms known to be related: anxiety, insomnia, and depression. We also analyzed the correlation between changes in these related symptoms.

The rate of occurrence of adverse events, hospitalizations, and emergency room visits were reported as absolute rates and percentages. Safety analyses comparing treatment groups on adverse events were conducted using a chi-squared test when at least 10% of participants reported an adverse event within a treatment group. Attrition rates and adequacy of treatment allocation concealment between groups were compared using chi-squared tests.


A total of 11,036 unique visits were made to the study website, and 1551 people initially registered for the study. Identity was verified for 1241 people (73%), of whom 540 (43%) met initial screening criteria. Of these, 449 (83%) returned for a second assessment of symptoms and 391 (87%) met eligible scores for anxiety and insomnia during the second assessment. All 391 participants were randomized to receive placebo (n = 135), kava (n = 121), or valerian (n = 135). Randomization and enrollment of the target sample size was completed within 8 weeks of study inception.

There were no statistically significant differences between the treatment groups' baseline clinical and demographic characteristics except that those randomized to placebo had higher scores on the initial screening instrument for depression (Table 1). Participants in the trial resided in 45 states including Hawaii and Alaska (Figure 1). Postbaseline data on outcomes were obtained for 116 (86%) of participants randomized to placebo, 105 (87%) of participants randomized to kava, and 122 (90%) of participants randomized to valerian. Study completion rates were similar across treatment groups (χ2 = 1.12; p = 0.57), with 108/135 (80%) participants completing placebo treatment, 97/121 (80%) completing kava treatment, and 114/135 (84%) completing valerian treatment. There were no significant differences between active and placebo groups in reasons for discontinuation of the study compounds (Figure 2).

Demographic and Clinical Characteristics of Randomized Patients at Baseline
Study recruitment by state (n = 391). Numbers represent the number of participants enrolled in the trial for each state. Participants were from 45 states including Hawaii and Alaska.
Trial flow diagram. The number of unique visits to the website is the number accrued during study enrollment. The number of participants registered is defined as the number of individuals who demonstrated interest in enrolling in the study by establishing an electronic signature and giving consent to provide their name, e-mail address, mailing address, and telephone number. Identity verification was conducted using electronic database services (Experian Inc., Orange, CA) to verify the accuracy of name and address information. Randomized participants were defined as not commencing therapy if they did not receive and sign for the study package despite multiple postal mail attempts or if after receipt of the study package they decided not to participate in the trial and returned the study compounds unopened. Postbaseline measures were obtained on all participants who provided 2-week or 4-week follow-up data. Those who completed trial are defined as those participants who provided 4-week follow-up data.


Participants in the placebo group had substantial reductions in anxiety (11.1 point decrease in STAI-State subtest score) after 2 weeks that continued into the fourth week of treatment (−14.4). Participants who received kava had a slightly smaller reduction in anxiety (−11.8; Table 2; Figure 3). The difference between groups was not statistically significant (decrease in kava − decrease in placebo = +2.6; 95% confidence interval [CI], −0.8 to +6.2). Participants receiving valerian had similar reductions in anxiety that were not statistically different from placebo (see Table 2).

Efficacy Measures*
Effects of kava or valerian on anxiety. Lower scores on the State-Trait Anxiety Inventory (STAI) are associated with fewer anxiety symptoms. The vertical bars depict 95% confidence intervals of the STAI mean score for each treatment arm at each study interval.

Participants reported substantial improvements in all indices of sleep 2 weeks after starting placebo, and these persisted into the fourth week (8.3 point decrease in ISI score). Those who received valerian had similar improvements (see Table 2; Figure 4). The difference between groups was not statistically significant for the change in ISI scores (decrease in valerian − decrease in placebo = +0.4 points; 95% CI, −1.3 to +2.1; see Figure 4). Participants receiving kava had similar reductions in insomnia scores that were not statistically different from placebo (see Table 2). Findings for anxiety and sleep measures remained unchanged after adjusting for baseline differences in depression scores.

Effects of kava or valerian on sleep. Lower scores on the Insomnia Severity Index (ISI) are associated with fewer insomnia symptoms. The vertical bars depict 95% confidence intervals of the ISI mean score for each treatment arm at each study interval.

Among adherent participants (n = 291; ≥80% of study compound taken), there were also no statistically significant differences between placebo and kava or placebo and valerian in the STAI-State subtest or the ISI (Table 3). Analyses of participants above and below the median score for the STAI-State subtest (median = 56) and the ISI (median = 16), and with and without a positive screen for depressive symptoms failed to identify statistically significant effects for the primary outcome measures for either active agent when compared with placebo (see Table 3).

Subgroup Analysis*

Treatment Adherence and Adequacy of Treatment Allocation Concealment

In the placebo group, 90% of participants took at least 80% of the capsules, compared with 85% of participants in the kava group and 82% of participants in the valerian group. There were no statistically significant differences between the 3 groups in the rates of adherence (χ2 = 3.5; p = 0.18).

Adequacy of treatment allocation concealment was assessed by asking participants if they had received active treatment or placebo. There were no statistically significant differences between groups in the proportion of participants who reported receiving active therapy (35% placebo, 33% kava, 43% valerian; χ2 = 2.9; p = 0.24).

Adverse Events

Adverse events occurred with similar frequency between active and placebo groups (Table 4). The only significant difference was a more frequent report of diarrhea among those receiving valerian (18%) compared with those receiving placebo (8%) (χ2 = 5.5; p = 0.02).

Adverse Events*

This study was conducted before the safety warnings about kava-related hepatotoxicity. Subsequent to these warnings, all 121 study participants receiving Kava were contacted by e-mail and postal mail, and none reported adverse events related to liver injury.

Internal Consistency of Participant Responses

There was a high correlation between depressive symptoms and STAI-State subtest score at week 4 (Spearman coefficient r = 0.72). We also observed significant correlations between change in ISI score and in STAI-State subtest score (Pearson coefficient r = 0.62).


To our knowledge, this trial is the largest randomized placebo-controlled trial of kava and valerian for anxiety and insomnia. We found that adults with at least mild anxiety, as assessed by the STAI-State subtest on 2 separate occasions, had substantial decreases in anxiety while taking placebo, but neither kava nor valerian improved symptoms of anxiety more than placebo. Similarly, adults who had at least mild insomnia for at least 2 weeks had substantial improvements in several indices of sleep (including improved ratings on the ISI, decreased sleep latency, and fewer nocturnal awakenings) with placebo treatment, but neither valerian nor kava further improved these indices of insomnia. These results suggest that people who feel anxious and experience insomnia, and seek self-treatment using these over-the-counter herbal compounds may often experience substantial improvement in symptoms during the first few weeks, but these improvements may not be attributed to the biological effects of kava or valerian.

Clinical trials for the treatment of anxiety requiring formal clinician-administered psychological testing have found similar placebo-group response rates (20%-25%)33,42. However, the placebo-group response rate for the sleep measures was higher than commonly seen in the literature12,32. These improvements may be attributed to the natural history of insomnia or a substantial physiological effect of participation in a clinical trial and receiving placebo capsules. We attempted to reduce the propensity for regression to the mean by excluding participants who had episodic insomnia and anxiety by requiring participants to meet inclusion criteria for both conditions on 2 separate occasions. As reported in a meta-analysis of the placebo effect, the physiological response to participating in a trial and receiving a placebo intervention can have beneficial effects on continuous subjective outcomes when compared with no treatment18.

As far as we know, this is the first report of a randomized, blinded trial conducted entirely via the Internet. Studies have documented high reliability between online and printed surveys4 and greater learning efficiency with online compared with the printed tutorials4. Researchers report successfully collected over 20,000 anonymous questionnaires on depression using the Centers for Epidemiological Studies Depression (CES-D) scale in an 8-month period using an Internet-delivered screening program17. Researchers using an Internet-delivered intervention study demonstrated successful subject participation and compliance, reliable data collection, and identification of an effective intervention for weight management43. The authors of a previous randomized controlled trial used the Internet to conduct several portions of the trial process including recruitment of participants and electronic data capture, although certain critical steps were performed using conventional methods29. The investigators required radiographs, informed consent, and medical record release forms to be mailed to the coordinating center to determine eligibility, and to enroll and randomize participants into the study. They concluded that these requirements impeded the speed of enrollment and suggested that methods to circumvent the need for paper documents, as implemented in the current study, would greatly improve efficiency. Similar to the current trial, that study estimated a 50% cost reduction compared with conventional trials. Incorporating methods from these studies, we developed a direct-to-participant method that allowed participants to enroll and participate in the trial using the Internet from the convenience of their home or work. This approach expands participation in clinical trials beyond geographic limits of proximity to clinical centers or medical offices. Trials conducted in this fashion can be completed relatively quickly: we registered and screened over 1500 potential participants and randomized 390 into the trial within 8 weeks.

We considered several potential explanations for the lack of significant difference between the herbal and placebo groups in the current trial. To our knowledge, this is the largest trial to date of the effectiveness of either herb and was designed with sufficient power to detect relatively small differences between groups. The confidence intervals around the observed results excluded clinically important effects. Adherence rates were relatively high, treatment allocation was adequately concealed, and there were no significant differences between the effects of the herbs and placebo among those with high levels of adherence. We also used validated instruments that have been used in other positive trials to assess severity of anxiety and insomnia, and we employed multiple measures to detect change in insomnia symptoms. In addition, we found variability in response that is similar to what has been observed in other studies of adults with insomnia or anxiety31,32,45. We considered the possibility that participants may provide less consistent responses using the Internet; however, we found a high correlation between improvements in anxiety and insomnia, suggesting participants were not intentionally entering random responses and providing misleading information. Although we did not take measures to verify identify through face-to-face contact with participants, we took several measures to avoid enrolling people who might give misleading information including obtaining confirmation of name and address using multiple databases, obtaining signature for delivery of study compounds, making phone contact with participants, and providing no monetary incentive to participate in the trial. Furthermore, face-to-face contact in conventional controlled trials does not preclude the provision of misleading or inaccurate information by participants. Our results may not be applicable to treatment of patients with generalized anxiety disorder in clinical practice, but are more applicable to the common self-administration of kava for relief of anxiety and valerian for the relief of insomnia.

It is possible that higher doses or longer periods of use of the herbals might have produced greater effects. However, the doses chosen in this trial are equal to or higher than those used in previous trials of kava and valerian, and the dose of active ingredients in the herbal capsules was confirmed by high-performance liquid chromatography assay. Additionally, there is no biological reason that more than 4 weeks of treatment would be required for these agents to have measurable effects. Specifically, 3 out of 5 trials that found an effect of kava6,23,44 and all 3 of the prior positive studies of valerian involved treatment for 4 weeks or less21,27,49.

Five of the 7 previous randomized, placebo-controlled trials of kava reported a reduction in anxiety symptoms in a standardized instrument6,23,44,48,50. Three of the 5 positive trials6,44,48 had 1 or more major methodologic problem, indicated by a Jadad score20 of less than 5. The Jadad evaluation system rates randomized controlled trials on a 0-5 scale and gives 1 point for each of the following: 1) the study was described as randomized, 2) the randomization technique was described and appropriate, 3) the study was described as double-blind, 4) the method of double-blinding was described and was appropriate, and 5) withdrawals and reasons for withdrawals were given. Specifically, these trials did not describe randomization procedures, did not include details on masking and blinding, and did not assess the adequacy of the treatment allocation concealment. Other problems included use of nonstandardized instruments and receipt of funding by the company providing the study compound.

A review of the previous randomized, placebo-controlled trials of valerian reveals less convincing evidence of efficacy. Of the 9 previous studies, only 321,27,49 reported positive results, and 610,21,24,26,27,41 of the 9 studies scored only 3 out of 5 on the Jadad scale. The sample sizes of the negative studies were extremely small, and most did not have significant power to identify clinically important benefits. None of the prior studies explicitly described any efforts to blind participants to the potent herbal smell of valerian, and therefore it is possible that previous positive effects were due to patients' awareness of the treatment they were taking. In contrast, in the current trial we performed explicit methods to mask odor, and adequacy of treatment-assignment concealment was demonstrated. Finally, it has been suggested that the literature about alternative and complementary treatment is particularly prone to publication bias, with negative trials remaining unpublished11,13.

In conclusion, we found that neither kava nor valerian was superior to placebo for the self-management of anxiety and sleep problems. This trial also demonstrates the feasibility of conducting rigorous randomized, blinded trials directly with participants entirely via the Internet.


We are indebted to the following for their contribution to this project: Michelle Arney for the development of the trial software; Skyli McAfee for directing participant recruitment. We acknowledge the assistance of Julian Nikolchev, Dan Thomson, David Karshmer, Susan Love, Tom McAfee, and Isaac Applbaum.


1. Ancoli-Israel S, Roth T. Characteristics of insomnia in the United States: results of the 1991 National Sleep Foundation Survey. I. Sleep. 1999;22(Suppl 2):S347-353.
2. Anonymous. Sleep in America. Princeton, NJ: Gallup Organization; 1995.
3. Balderer G, Borbely AA. Effect of valerian on human sleep. Psychopharmacology. 1985;87:406-409.
4. Bell D, Fonarow G, Hays R, Mangione C. Self-study from web-based and printed guideline materials. A randomized, controlled trial among resident physicians. Ann Intern Med. 2000;132:938-946.
5. Bell D, Kahn CJ. Health status assessment via the World Wide Web. Proc AMIA Annu Fall Symp. 1996:338-342.
6. Bhate H, Gerster G, Gracza E. Orale pramedikation mit zubereitungen aus Piper methysticum bei operativen eingriffen in epiduralanasthesie. Erfahrungsheilkunde. 1989;6:339-345.
7. Blumenthal M. The Complete German Commission E Monographs: Therapeutic Guide to Herbal Medicines. Austin, TX: American Botanical Council and Integrative Medicine Communications; 1998.
8. Blumenthal M. Herb sales down 3% in mass market retail stores-sales in natural food stores still growing, but at lower rate. HerbalGram. 1999;68-69.
9. Center for Science in the Public Interest. Caffeine content of food and drugs. Nutrition Action Newsletter. December 1996.
10. Donath F, Quispe S, Diefenbach K, Maurer A, Fietze I, Roots I. Critical evaluation of the effect of valerian extract on sleep structure and sleep quality. [Comment in: Pharmacopsychiatry. 2000;33:239.] Pharmacopsychiatry. 2000;33:47-53.
11. Easterbrook P, Berlin J, Gopalan R, Matthews D. Publication bias in clinical research. Lancet. 1991;337:867-872.
12. Edinger J, Wohlgemuth W, Radtke R, Marsh G, Quillian R. Cognitive behavioral therapy for treatment of chronic primary insomnia: a randomized controlled trial. JAMA. 2001;285:1856-1864.
13. Ernst E, Pittler MH. Alternative therapy bias [letter]. Nature. 1997; 385:480.
14. Fifer SK, Mathias SD, Patrick DL, Mazonson PD, Lubeck DP, Buesching DP. Untreated anxiety among adult primary care patients in a health maintenance organization. Arch Gen Psychiatry. 1994;51:740-750.
15. Foley DJ, Monjan AA, Brown SL, Simonsick EM, Wallace RB, Blazer DG. Sleep complaints among elderly persons: an epidemiologic study of three communities. Sleep. 1995;18:425-432.
16. Ford DE, Kamerow DB. Epidemiologic study of sleep disturbances and psychiatric disorders. An opportunity for prevention? [see comments]. JAMA. 1989;262:1479-1484.
17. Houston TK, Cooper LA, Vu HT, Kahn J, Toser J, Ford DE. Screening the public for depression through the Internet. Psychiatr Serv. 2001;52:362-367.
18. Hrobjartsson A, Gotzsche P. Is the placebo powerless: an analysis of clinical trials comparing placebo with no treatment. N Engl J Med. 2001;344:1594-1602.
19. Isacc RA. Plants. In: Cunniff P, ed. Official Methods of Analysis of AOAC International. 16th ed. Arlington, VA: AOAC International; 1995:1-35.
20. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials. 1996;17:1-12.
21. Kamm-Kohl AV, Jansen W, Brockmann P. Moderne Baldriantherapie gegen nervose storungen im senium. Medwelt. 1984;35:1450-1454.
22. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, Wittchen HU, Kendler KS. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States. Results from the National Comorbidity Survey. Arch Gen Psychiatry. 1994;51:8-19.
23. Kinzler E, Kromer J, Lehmann E. [Effect of a special kava extract in patients with anxiety-, tension-, and excitation states of non-psychotic genesis. Double blind study with placebos over 4 weeks]. Arzneimittelforschung. 1991;41:584-588.
24. Kuhlmann J, Berger W, Podzuweit H, Schmidt U. The influence of valerian treatment on "reaction time, alertness and concentration" in volunteers. Pharmacopsychiatry. 1999;32:235-241.
25. Kuppermann M, Lubeck DP, Mazonson PD, Patrick DL, Stewart AL, Buesching DP, Fifer SK. Sleep problems and their correlates in a working population. J Gen Intern Med. 1995;10:25-32.
26. Leathwood PD, Chauffard F. Aqueous extract of valerian reduces latency to fall asleep in man. Planta Medica. 1985;53:144-148.
27. Leathwood PD, Chauffard F, Heck E, Munoz-Box R. Aqueous extract of valerian root (Valeriana officinalis L.) improves sleep quality in man. Pharmacol Biochem Behav. 1982;17:65-71.
28. Lehmann E. Wirkung von Kava: Kava bei akuter Angst. Synopsis. 1998;2:59-64.
29. McAlindon T, Formica M, Kabbara K, LaValley M, Lehmer M. Conducting clinical trials over the internet: feasibility study. BMJ. 2003; 327:484-487.
30. Mellinger GD, Balter MB, Uhlenhuth EH. Insomnia and its treatment. Prevalence and correlates. Arch Gen Psychiatry. 1985;42:225-232.
31. Morin C. Insomnia: psychological assessment and management. New York: Guilford; 1993.
32. Morin CM, Colecchi C, Stone J, Sood R, Brink D. Behavioral and pharmacological therapies for late-life insomnia: a randomized controlled trial [see comments]. JAMA. 1999;281:991-999.
33. Piercy M, Saramek J, Kurtz N, Cutler N. Placebo response in anxiety disorders. Ann Pharmacotherapy. 1996;30:1013-1019.
34. Pittler MH, Ernst E. Efficacy of kava extract for treating anxiety: systematic review and meta-analysis. J Clin Psychopharmacology. 2000;20:84-89.
35. Rice DP, Miller LS. Health economics and cost implications of anxiety and other mental disorders in the United States. Br J Psychiatry Suppl. 1998;51:4-9.
36. Richman A, Witkowski J. 4th Annual Herbal Products Sales Survey. Whole Foods; 1998:19-26.
37. Richman A, Witkowski J. 6th Annual Herbal Products Sales Survey. Whole Foods; 2000:52-59.
38. Rickels K, Schweizer E. The spectrum of generalised anxiety in clinical practice: the role of short-term, intermittent treatment. Br J Psychiatry Suppl. 1998;51:49-54.
39. Rogers MP, White K, Warshaw MG, Yonkers KA, Rodriguez-Villa F, Chang G, Keller MB. Prevalence of medical illness in patients with anxiety disorders. Int J Psychiatry Med. 1994;24:83-96.
40. Sateia MJ, Doghramji K, Hauri PJ, Morin CM. Evaluation of chronic insomnia. An American Academy of Sleep Medicine review. Sleep. 2000;23:243-308.
41. Schulz H, Stolz C, Muller J. The effect of valerian extract on sleep polygraphy in poor sleepers: a pilot study. Pharmacopsychiatry. 1994;27:147-151.
42. Schweizer E, Rickels K. Placebo response in generalized anxiety: its effect on the outcome of clinical trials. J Clin Psychiatry. 1997; 58(Suppl):30-38.
43. Shochat T, Umphress J, Israel AG, Ancoli-Israel S. Insomnia in primary care patients. Sleep. 1999;22(Suppl 2):S359-365.
44. Singh NN, Ellis CR, Sharp I, Eakin K, Best AM, Singh YN. A 4-week, randomized, double-blind, placebo-controlled study to investigate the effectiveness and safety of Kavatrol was conducted in a non-clinical sample with anxiety and stress. Natrol; 1997.
45. Spielberger C. Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA: Consulting Psychologists Press; 1983.
46. Stevinson C, Ernst E. Valerian for insomnia: a systematic review of randomized clinical trials. Sleep Med. 2000;1:91-99.
47. Tate D, Wing R, Winett R. Using Internet technology to deliver a behavioral weight loss program. JAMA. 2001;285:1172-1177.
48. Volz HP, Kieser M. Kava-kava extract WS 1490 versus placebo in anxiety disorders-a randomized placebo-controlled 25-week outpatient trial. Pharmacopsychiatry. 1997;30:1-5.
49. Vorbach EU. Therapie von insomnien: wirksamkeit und vertraglichkeit eines Baldrianpraparats. Psychopharmakotherapie. 1996;3:109-115.
50. Warnecke G. [Psychosomatic dysfunctions in the female climacteric. Clinical effectiveness and tolerance of Kava Extract WS 1490]. Fortschr Med. 1991;109:119-122.
51. Warnecke G, Pfaender H, Gerster G, Gracza E. Wirksamkeit von Kawa-Kawa-Extrakt beim klimakterischen Syndrom. Z Phytother. 1990;11:81-86.
52. Whooley MA, Avins AL, Miranda J, Browner WS. Case-finding instruments for depression. Two questions are as good as many. J Gen Intern Med. 1997;12:439-445.
© 2005 Lippincott Williams & Wilkins, Inc.