No Effect of Lunar Cycle on Psychiatric Admissions or ...badmoonrising.wikispaces.com/file/view/No+Effect+of+Lunar+Cycle+o… · No Effect of Lunar Cycle on Psychiatric Admissions

MILITARY MEDICINE. 171. 12:1239.2006

No Effect of Lunar Cycle on Psychiatric Admissions orEmergency Evaluations

Guarantor: LCDR Robert N. McLay. MC USNRContributors: LCDR Robert N. McLay. MC USNR*; LCDR Amado A. Daylo, MC USNR+;CDR Paul S. Hammer. MC USNR|

It is a popularly held belief that psychiatric behavior worsensduring a full moon. Research in this area has yielded mixedresults. Records from Naval Medical Center San Diego for1993-2001 were examined to see whether there were higherrates of psychiatric admission associated with particularphases of the moon. Records from 8.473 admissions revealedthat there were no more admission on days with a full moon, anew moon, any quarter of the moon, a waxing moon, or a waningmoon. This held true for psychiatric patients as a whole, as wellas for individuals with particular diagnoses, such as those with amood disorder or psychotic disorder. Records from 1.909 emer-gency psychiatric evaluations that occurred between 2002 and2003 were also examined to see whether a higher percentage ofpatients might present, but not require hospitalization, during aparticular phase of the moon. Once again, no significant effectwas found. In summary, lunar phase was not associated in anysignificant way with psychiatric admissions or emergency pre-sentation.

Introduction

I t is a commonly held belief that the phases of the mooninfluence behavior. Surveyed emergency room personnel have

indicated the conWction that the demand for services increasesduring the full moon.' and several previous investigations haveattempted to link psychiatric emergencies with lunar cycles. Themajority of these studies have been negative.̂ "^ but sporadicreports have indicated a link between lunar phases and someforms of psychiatric event.'''''

There are three basic possibilities to explain the difference inresults, The negative studies could have had insufficient powerto detect differences. The positive studies could be reportingspurious relations of chance, or there eould be a particularsubset of the population that responds to the moon that showsup in some studies but not in others.

Those studies that have shown significant effects have oftendone so only in subset of the patient population, such as thosewith psychosis.'' or using a complicated pattern of lunar predic-tion, such as finding that patient census increases during thelast quarter moon.̂ Sinee there was no a priori reason to look forthat particular pattern, it is possible that so many possibilitieswere investigated that one came out looking significant. There

'Psychiatry Staff. Department of Mental Health, Naval Medical Center San Diego.3 4 ^ 0 Bob Wilson Drive. San Diego. CA 92123.

^Psychiatry Staff Physician. U.S. Naval Hospital, Yokosuka. Japan.IPsychiatry Staff Physician and Residency Director. Naval Medical Center. San

Diego,'CA 92123.The \iews expressed in this article are those of the authors and do not reflect the

official policy or position of the Department of the Navy. Department of Defense, or theU.S. govemment.

This manuscript was received for review in February 2005. The revised manuscriptwas accepted for puhlicalion in Januar\' 2006.

also may be a bias in the literature in that negative studies aregenerally more difficult to publish than those showing a result.It is telling in this regard that few studies show an effect of asingle lunar phase on a particular behavior. The exception tothis is that more than one study has demonstrated an increasein acting out or aggressive behavior around the time of the fullmoon.^''" Even in this particular area, however, there are morerobust studies arguing against a lunar effect."'^

A meta-analysis by Rotton and Kelly'̂ in 1985 showed overallthere was no effect of lunar cycle. This has yet to be confirmed bya large-scale independent study, however. Also, no subsequent,large-scale study has examined whether the lunar cycle mighthave an effect on one particular subset of psychiatric diagnosis.Medical records from Naval Medical Center San Diego (NMCSD)offered the opportunity to do just this. A data set. which hadbeen previously used to examine length of stay.'"" showed datesof admission for all psychiatric patients from September 23.1993 to June 18, 2001. This included information on 8.473psychiatric admissions, and provided the largest study yet pub-lished on the effects of lunar cycle on psychiatric admissions.Unfortunately, from this dataset. information was available onlyon individuals requiring hospitalization. not those who came infor emergent evaluation of less severe problems. This is impor-tant to examine because in pre\1ous work.'^ it has been shownthat it is these "worried well" that appear to be influenced bysubtle factors such as the weather.

This problem was addressed by examining such an effect in1.909 patients who came through the emergency room (ER)from August 1, 2002 to July 31. 2003. This provided an oppor-tunity to look at not only patients who were admitted to thehospital, but also those who were evaluated in the ER andreleased. These two data allowed a comprehensive look at theeffects of lunar cycle on psychiatric presentation.

Methods

Data Set 1 (Psychiatric Admissions)

Dates of admission were noted for all psychiatric patientsfrom September 23. 1993 to June 18. 2001. This included in-formation on 8,473 separate admissions. Previous use of thisdata showed that this included 6,366 unique, first-time admis-sions.''* likely representing that same number of individual pa-tients. Although no indi\1dual tracking information was avail-able, data were available on primary diagnosis associated witheach admission. Admissions were divided by Diagnostic andStatistical Manual criteria into the categories of those with psy-chotic disorders (Admits Psychotic), mood disorders (AdmitsMood), anxiety disorders (Admits Anxiety), and other psychiatricdiagnoses (Admits Other). A total of all admissions each day was

1239 Militaiy Medicine. Vol. 171. December 2006

1240 No Lunar Cycle Effect on Psychiatric Admissions

TABLE I

ADMISSIONS PER DAY FROM SEPTEMBER 23. 1993TO JUNE 18,2001

TABLEm

PSYCHIATRIC ADMISSIONS PER DAY DIVIDED BY LUNAR QUARTER

T\pe of admissionPsychotic-Mood disorderAnxiety disorderOther disorderTotal

Minimum

00000

Maximum

3528

13

Mean

0.230.860.131.783.00

SEM

0.010.020.010.030.04

SD

0.490.970.371,431.95

also noted (Admits Total). Date of occurrence was noted for eachtype of admission so that numbers could he compared to thelunar cycle on that date. Descriptive statistics (including themean, standard deviation (SD), and standard error of the mean(SEM)) for this data were reported in Table I.

Data Set 2 (Psychiatric Emergency Evaluations)

Data for emergency psychiatric visits at the NMCSD wereobtained from a 1 -year period. August 1. 2002 to July 31. 2003.The data were compiled from a computerized log created forrecording all of the consult-liaison and on-call psychiatric con-sultations periormed by mental health services at the hospital.This information was subdivided according to evaluations thatrequired admission, and those released from the ER. The per-centage of evaluated patients who were admitted each day (%Admit) was recorded each day and matched to lunar cycle.Descriptive statistics (including the mean. SD. and SEM) for thisdata were reported in Table II.

Lunar Cycle

Information on lunar cycle was obtained from the U.S, NavalObservatory. This gave the percentage of the moon illuminatedin San Diego. California, for all of the dates covered hy thisstudy. For purposes of this study, a moon illumination of 100%was considered a full moon, and an illumination of <1% wasconsidered a new moon. Lunar cycle was diWded into quarters.with each 25% illumination accounting for the flrst, second,third, and fourth quari:er, respectively.

Statistics

Differences in the numher of psychiatric admissions and eval-uations were compared throughout the lunar cycle. The Studentt test was used to evaluate possible differences between dayswith a full moon and those without, days with a new moon anddays without, and waxing and waning moons. Analysis of vari-ance (ANOVA) was used to examine potential differences amongthe four quarters of the moon. From the first data set. outcome

TABLEn

EMERGENCY PSYCHIATRIC EVALUATIONS FROM AUGUST 1. 2002TO JULY 31.2003

Total evaluationsAdmittedReleased% Admission

Minimum

0000

Maximum

188

14100

Mean

5.131.523.61

28.9

SEM

0.160.080.131.3

SD

3.071.492.42

24.6

Lunar Quarter

FirstMeanMinimumMaximumSEM

SecondMeanMininiimiMaximumSEM

ThirdMeanMinimumMaximumSEM

FourthMeanMinimumMaximumSEM

Psychotic

0.230.003.000.02

0.220.003.000.02

0.230.002.000.02

0.230.003.000.02

Type of Admission

Mood

0.850.005.000.03

0.880.005.000.04

0.880.005.000.04

0.850.005.000.03

Anxiety

0.130.002.000.01

0.150.002.000.02

0.110.002.000.02

0.140.002.000.01

Other

1.860.008.000.05

1.740.007.000.07

1.760.008.000.07

1.720.008.000.05

Total

3.080.00

13.000.07

3.000.009.000.08

2.980.009.000.09

2.940.00

10.000.06

measures were the following: Admits Psychotic, Admits Mood,Admits Anxiety. Admits Other, and Admits Total. From the sec-ond data set. % Admit was used as the outcome variable. Posthoc power analysis was used to compute the power of this studyto detect a difference of a single admission or evaluation per dayin all t test evaluations.

TABLE IV

PSYCHIATRIC ADMISSIONS PER DAY, SHOWN BY TYPE OF MOON

Lunar Phase

Full moonMeanMinimumMaximumSEM

New moonMeanMinimumMaximumSEM

Other moonMeanMinimumMaximumSEM

Waning moonMeanMinimumMaximumSEM

Waxing moonMeanMinimumMaximumSEM

Psychotic

0.210.002.000.04

0.290.002.000.05

0.230.003.000.01

0.220.003.000.01

0.230.003.000.01

Type of Admission

Mood

0.880.003.000.08

0.910.005.000.09

0.860.005.000.02

0.870.005.000.03

0.850.005.000.03

Anxiety

0.150.002.000.04

0.110.001.000.03

0.130.002.000.01

0.140.002.000.01

0.120.002.000.01

Other

1.740.006.000.13

1.830.006.000.13

1.780.008.000.03

1.790.008.000.04

1.760.008.000.04

Total

2.970.007.000.16

3.140.008.000.16

3.000.00

13.000.04

3.040.00

13.000.05

2.970.00

11.000.05

Militaiy Medicine. Vol. 171, December 2006

No Lunar Cycle Effect on Psychiatric Admissions 1241

TABLE V

EMERGENCY PSYCHIATRY EVALUATIONS PER DAY AND THEPERCENT THAT RESULTED IN ADMISSION

Lunar Quarter

FirstMeanMinimumMaximumSEM

SecondMeanMinimumMaximumSEM

ThirdMeanMinimumMaximumSEM

FourthMeanMinimumMaximumSEM

Evaluations

5.220.00

18.000.29

5.370.00

16.000.40

5.020.00

15.000.40

4.950.00

14.000.26

Admitted

1.400.006.000.13

1.870.008.000.21

1.780.007.000.24

1.320.005.00O.ll

Released

3.820.00

14.000.24

3.500.00

11.000.28

3.240.00

12.000.29

3.630.00

13.000.22

% Admit

26.80.0

100.02.3

32.50.0

100.02.9

32.80.0

lOO.O3.6

27.20.0

100.02.2

Results

Mean numbers of psychiatric admissions from Septemher 23,1993 to June 18, 2001. grouped by quarter of the moon anddiagnosis, are given in Table III. Similar information grouped bywhether there was a full or new moon, and whether the moon

TABLE VI

PERCENTAGE OF PSYCHIATRIC EVALUATIONS REguiRlNGADMISSION BY LUNAR PHASE

Lunar Phase

Full moonMeanMinimumMaximumSEM

New moonMeanMinimumMaximumSEM

Other moonMeanMinimumMaximumSEM

Waning moonMeanMinimumMaximumSEM

Waxing moonMeanMinimumMaximumSEM

Evaluations

4.140.008.000.62

4.822.00

10.000.59

5.180.00

18.000.17

5.130.00

18.000.16

5.360.00

16.000.24

Admitted

t.OO0.003.000.28

1.290.003.000.27

1.550.008.000.08

1.520.008.000.08

1.630.008.000.12

Released

3.140.006.000.47

3.531,007.000.45

3.630.00

14.000.14

3.610.00

14.000.13

3.730,00

13.000.19

% Admit

22.10.0

66.76.2

25.30.0

66.75.1

29.30.0

100.01.4

28.90.0

100.01.3

30.30.0

100.01.9

was waxing or waning, is given in Table IV. No statisticallysignificant differences related to quarter of the moon were seenfor Admits Psychotic (p = 0.99}. Admits Mood [p = 0.87), AdmitsAnxiety (p = 0.30). Admits Other (p = 0.17), or Admits Total {p =0.50). Also, no statistically significant differences were seen forany of these variables when comparing days with a full moon tothose without (p = 0.54,0.81.0,57,0,74, and 0.87. respectively)or days with a new moon and those without (p = 0,23, 0.60,0.38. 0.69, and 0.40, respectively). There was also no differencefor these variables depending on whether the moon was waxingor waning (p = 0.53. 0.45, 0,12. 0.54. and 0.34, respectively).

Similarly, in looking at patients seen for emergency psychiat-ric evaluation between August 1, 2002 to July 31, 2003, nostatistically significant differences were associated with lunarcycle. Comparing days with a full moon (p ^ 0,71). a new moon(p = 0.49), among quarters of the moon (p = 0,26). or between awaxing and waning moon (p = 0,29). there were no statisticallysignificant differences in the percentage of emergency psychiat-ric evaluations that required psychiatric admission. Mean num-bers of emergency psychiatric evaluations, admitted and re-leased, grouped by lunar quarter, are given in Table V. In TableVI. similar information is shown for days with a full moon, newmoon, other phase, and if the moon was waxing or waning.

Post hoc power analysis revealed that given the size of thisstudy, and the SD discovered, power to detect a difference ofasingle admission per day between days with full moons andother types of moons was >99% for Admits Total, and admis-sions of each diagnosis type. Conversely, if power was set at 80%this study would have had the ability to detect a A of 0,5 AdmitsTotal. 0.1 Admits Psychotic, 0,3 Admits Mood, 0.1 Admits Anx-iety, or 0.4 Admits Other. For the smaller database examining %Admit, post hoc power analysis showed that with a power of80%, this study would have been able to detect a A of 20%between days with full moons and those without.

Discussion

No statistically significant effect of the lunar cycle was seen onpsychiatric admissions or evaluations of any Idnd. There wereno more psychiatric patients admitted on days with a full moonor days with a new moon. When comparing the four quarters ofthe moon, no differences were seen, A waxing or waning moonmade no difference. There was still no effect of lunar cycle evenif the data were broken down by diagnosis. Psychotic, mood-disordered, and anxiety-disordered patients all ended up in thehospital at a rate independent of the phases of the moon.

This study was sufficiently large that it should have beenpossible to detect even subtle differences in rates of psychiatricadmission on days with a full moon. The power of this study todetect small differences in rates of evaluations not requiringhospitalization was less robust. However, the overall pattem ofadmission rates did not support the idea of doing a larger studyon nonhospitalized patients. Frequently, evaluations during afull moon were, if anything, less frequent than during otherlunar phases.

It should be noted that in examining so many potential waysthat the moon might have an effect, this study was predisposedtoward type I error. That is. because of the large number ofoutcome variables, there was a chance that one of them couldhave spuriously shown what appeared to be a statistically sig-

Miiitary Medicine. Vol. 171. December 2006

1242 No Lunar Cycle Effect on Psychiatric Admissions

nificant difference. The absence of a detectable effect in such astudy argues all the more strongly, however, that the moon isnot influencing psychiatric presentation.

Our study did not specifically examine aggression, but onewould expect that if there were an increase in violence aroundthe time ofthe full moon, this would have resulted in a greaternumber of ER evaluations or psychiatric admissions, especiallyat a hospital that serves a large population of young, malesailors and marines.

Clearly, despite the evidence against it. the idea that the lunarcycle influences behavior is popular, including among ER' staffand mental health workers."' This is not to say that there is noconceivable way in which lunar cycle alters behavior. For exam-ple, the human behavior of uttering "my what a nice, full moon"is almost assuredly associated with the actual lunar phenom-ena. There has been some debate as to whether dog bites in-crease during a full moon,'''''^ Also, it has been hypothesizedthat there once was a more dramatic effect of lunar cycle onhuman behavior, hut that the invention of the electric lightdrowned the luminous powers of the moon."^ Regardless, ifthese effects do or did exist, this study, and the majority of otherevidence, would argue that the moon is not influencing theworkload of psychiatric staff. Therefore, contrary to the conten-tions of those medical providers surveyed by Danzel' and others,adjusting schedules or pay according to lunar cycle makes littlesense. In short, a big moon makes no big difference.

References

1. Danzel DF: Lunacy. J Emerg Med 1987: 5: 91-5.2. Bauer SF, Homick EJ: Luuar elTect on mental illness: Ihe relationship of tnoon

phase to psychiatric emergencies. Am J Psychiatry 1968; 125: 696-7.

3. Gorviti JJ . Roberts MS: Ltiiiar phases and psychiatric hospital admissions. Psy-cho! Rep 1994: 75: 1435-40.

4. Amaddeo F. BisoITi G. Micciolo R, Piccinelli M. Tansella M: Frequeticy of contactwilh cotnmunily-based psychiatrir servires and the lunar cyrle: a 10 year case-register study. Soo Psychiatry Psychiatr Epidemiol 1997: 32: 323-6.

5. Geller SH. Shannon HW: The moon, weather and tnental hospital contacts:confimiation and explanation of the Transylvania effert.,} Psychiatr Nurs MentHealth Sen' 1976: 14: 13-7.

6. Templer DI, Veleber DM: The moon and madness: a comprehensive perspective.J Clin Psychol 1980: 36: 865-8.

7. Barr W: Lunacy revisited: the influence of the moon on mental health and qualityor life. J Psychosoc Nurs Ment Health Serv 2000: 38: 28-35.

8. Temte JL: Exploring environmentai cycles in psychiatric patients. WMJ 1989:86:17-20.

9. Hicks-Caskey WE, Potter DR: Effect of the full moon on a sample of developmen-tally delayed, institutionalized women. Percept Mot Skills 1991; 72: 1375-80.

10. Lieher AL: Human aggression and the lunar synodic cycle. J Clin Psychiatry1978: 39: 385-92.

11. Nunez S. Pereic Mendez L. Agtiirre Jaime A: Moon cycles and violent behaviours:myth or fact? Eur J Emerg Med 2002: 9: 127-30.

12. Owen C. Tarantello C. Jones M. Tennant C: Lunar cycles and violent behaviour.Aust N Z J Psychiatry 1998: 32: 496-9.

13. Rotton J, Kelly IW: Much ado aboul the full moon: a meta-analysis of lunar-lunacy research. Psychol Bull 1985: 95: 286-306.

14. McLay RN, Daylo A. Hammer P: Predictors of length of stay in a psychiatric wardserving active duty military and civilian patients. Milit Med 2005: 170: 219-22.

15. McLay RN. Santiago PN, Daylo A. Hammer P: Effects of Weather and Lunar Cycleon Psychiatric Emergency Evaluations. Abstract. 2005 American PsychiatricAssociation Annual Meeting. May 23, 2005, Atianta, GA.

16. Vance DE: Belief in lunar effects on human behavior. Psyrhol Rep 1995: 76: 32-4.17. Bhattachar)ee C. Bradley P. Smith M, Scally AJ. Wilson BJ: Do animals bite

more duringa full moon: retrospective observational analysis. BMJ2000: 321:1559-61.

18. Chapman S. Morrell S: Barking mad: another lunatic hypothesis bites the dust.BMJ 2000: 321: 1561-3.

19. Raison CL. Klein HM. Steckler M: The moon and madness reconsidered. J AffectDisord 1999: 53:99-106.

If someone iieoi you collapses, loses consciousness and stops breathing, chances

are their heart has stopped. Call 9-1-1 immediately and do CPR until help arrives.

If you don't know CPR, learn. To find a ^

CPR course near you, call the American American HeartHeart Association at 1-877-AHA-4CPR or Associarion.

visit us online at www.amer(canheart.org. Learn and Live.T)i« ipaoa piovKtad SB B pjCbc $erv>c« O 2D00. Ammcan Hean AsioOBlton

Military Medicine. Vol. 171, Deeember 2006