Based on an experiment with a small sample size and the observations of improved symptoms with lavender pillows after stopping medications, the author concludes that lavender cures insomnia. The study design is flawed and the methodology is insufficient due to lack of control group, randomizations, small sample size, unclear metrics for measuring the quality of sleep and unknown long term effect, which renders the conclusion unwarranted.
One of the salient issue is the lack of control group, since everyone is exposed to lavender, no real control group with chronic insomnia is identified to compare the results. The author could benefit from setting up either blank control who has been diagnosed with chronic insomnia without any medications or active controls who is taking medications while controlling for other factors and compared the results to see whether exposure to lavender could benefit sleep. The author might attempt to do a cross over study with every one serving as its own control. If this is case, the study still need more evidence to ensure the validity of the design, since by exposing the study participants to lavender pillow together with medications, the author risk contaminating the effects of lavender, unless he could prove that lavender will not take effect until 2 weeks later,and the first week’s effect can be simply be attributed to medication alone. A more rigorous design is to expose participants to medications only in the first week following by a washout period and then exposing to lavender in the final week to avoid the mixed effects. This actually also raises another related point: The effects of medication might be long-lasting and lasts until week 2 or even week 3, making it hard to distinguish between lavender and meds effects. Evidence about the half life of medication and washout period duration is needed in order to separate lavender effects from medication. Moreover, the improvement of sleepiness could be explained by other reasons with the flawed design , for example, patients might less likely to sleep well in the new environment while after 2 weeks, they might get used to the new environment , leading to the improved quality of sleep, which has nothing to do with either lavender or medication.
All of the aforementioned counterarguments are based on the assumption that the results are correct and the measurement of sleep quality is accurate, however, a closer examination of the results will find that the sleep quality metric is not well defined. What does it mean when the author mentions ‘sleep more soundly”? Is it refers to the decreased time to sleep, increased rapid eye movement, and reduced restless time during sleep? More information is needed. Also, longer sleep duration may not be statically or clinically meaningful and it still be under the normal ranges for people without chronic fatigue. Moreover, there is also a lack of specific numbers , proportions and statistical testing regarding the improved sleep and whether the improvement is significant both clinically and statistically. Based on the small sample size of 30 participants, it remains unlikely. The small sample size, of course, also raise another issue of generalizability, more evidence is needed regarding those volunteer share the same demographic characteristics in terms of age, sex, comorbidities and how generalizable the study results are to the whole population. Three week time period is also too short to reach the conclusion that lavender CURES insomnia since the long term effect still remains unknown and the effect of improved sleep ( if warranted) could still be temporary. Evidence regarding whether relapse is possible after three weeks is needed to prove lavender CURES insomnia, especially given the ‘chronic’ nature, if relapse does occur, the conclusion could be weakened.
In summary, the study is flawed in design due to lack of real control group, contaminated effects between medications and lavender and insufficient definitions of sleep outcome and specific numbers, proportions, statistical tests results and long term effects render the results unconvincing