Interaction between bedding and sleeping position in the sudden infant death syndrome: a population based case-control study

Please download to get full document.

View again

of 5
6 views
PDF
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Document Description
Interaction between bedding and sleeping position in the sudden infant death syndrome: a population based case-control study
Document Share
Document Tags
Document Transcript
  indicate thatattitudes among the medical profession and patientsare changing. There has been much discussion about clinical trials andinformed consent,46 and, though we believe that informed consent is mandatory, we accept thata proportion of patients do not enter trials becausethey may dislike the idea of a random decision being made about treatment or because they refuse or request one part of the random option offered. A further group of patients are excluded because they do not fulfil theentry criteria for the trial, and this group may be larger than forecast at the time the trial was planned. These factors may be the reason why accrual to the Scottishbreast conservation trial has been slower than anticipated. The planned total intake was 900 patients, and after four years 420 patients hadbeen entered. In conclusion more than half of the patients thought initially to be suitablefor conservation were excluded from our trials and one third ofthe remainder refused to take part. Those planning prospective clinical trials should thereforetake into account the loss of patients through ineligibility and refusal when predicting theaccrual rate and overall duration of anyproposed trial and thepossible effects of this selection on conclusions drawnfrom the results. I De Vita V-r. Breastcancer therapy: exercising all our options.   EnglJI Med 1989;320:527-9. 2 Stewart HJ, Prescott RJ, Forrest APM. Conservationtherapy ofbreast cancer. Lancet 1989;ii:168-9. 3 Wilson RG, HartA, Dawes PJDK. Mastectomy or conservation: thepatient'schoice. BrMedJ, 1988;297:1167-9. 4 Brahams D.Medicine and thelaw. Randomised trials and informed consent. Lancet 1988;ii: 1033-4. 5 Baum Mi, Zilkha K, Houghton J. Ethicsof clinical research:lessons forthe future. BrMedj 1989;299:251-3. 6 Brewin CR, Bradlev C. Patient preferences andrandomised clinical trials. BrMed7 1989;299:313-5. (Accepted 25 April 1990) Interaction between bedding and sleepingposition in the sudden infant death syndrome: a population based case-control study Peter J Fleming, Ruth Gilbert, Yehu Azaz, P Jeremy Berry, Peter T Rudd, Alison Stewart, Elizabeth Hall Abstract Objective-To determine the relation between sleepingposition and quantity of beddingand the risk of sudden unexpected infant death. Design-A study of all infants dyingsuddenly and unexpectedly and of two controls matched for age and date with each index case. The parents of control infants were interviewed within 72 hours of the index infant's death. Information was collected on bedding, sleepingposition,heating, and recent signs of illness for index and control infants. Setting-A defined geographical area comprising most of the county of Avon and part of Somerset. Subjects-72 Infants who had died suddenlyandunexpectedly (of whom 67 had died from the sudden infant death syndrome)and 144 control infants. Results-Compared with the control infants the infants who had died from the sudden infant death syndrome were more likelyto have been sleeping prone (relativerisk8-8; 95 confidence interval 7.0to 11-0; p<0-001), to havebeen more heavily wrapped (relativerisk 1-14 per tog above 8 tog; 1-03to 1-28; p<005), and to have had the heating on all night (relativerisk2-7; 1-4to 5.2; p<0-01). These differ- ences were less pronounced in the younger infants (less than 70 days)than the older ones. The risk of suddenunexpected death among infants older than 70 days, nursed prone, and with clothing andbedding of total thermal resistance greater than 10 tog was increased by factors of 15-1 (2.6 to 89.6) and 25-2  3.7 to 169-0) respectively compared with the risk in infants of the same age nursed supine or on their side andunder less than 6tog of bedding.Conclusions -Overheating and the prone position are independently associated withanincreased risk of suddenunexpected infant death, particularly in infants aged more than70 days. Educating parents about appropriatethermal care and sleeping position of infants may help to reduce the incidenceof the sudden infant death syndrome. Introduction The possible role of thermal stress in theaetiology of the sudden infant death syndrome has been suggested by many authors,' 4 andWailoo et al recently showed that many babies are put to bedunder excessive amounts of bedding.4 The low incidence of the sudden infant death syndrome in Hong Kong, which has a hot, humid climate but wheremost infantssleep supine, has led to the suggestion that the prone position may be an important risk factor.4 Recently, Nelson et al suggested on the basis of a simple model of infantheat balance that infants sleeping in the prone position with an excess of bedding would be more likely to become hyperthermic than infantsin the supine position with equal bedding.6 We have shown an appreciable rise in oxygen consumptionand carbon dioxide production (and hence heat production) between birth and   month of age in healthy infants.7 This highermetabolic rate is maintained until at least 3 months. We have also shown that raising the environmental temperature around healthy infants aged 3 months or less increases respiratory oscillations, suggesting an effect on the respiratorycontrolsystem.' Thus infants in this age range, which is the age of peak incidence of the sudden infant death syndrome, would be at increased risk of the consequences of overwrapping. The effects of overwrapping would be likely to be greater at the time of acute viral infection, when the metabolic rate rises, and two studies have shown that many parents in the United Kingdom and in New Zealand respond to infections in their babies by increasing the amount of clothing and bedding.90 To investigatethepossibleinteractions between quantity of bedding and sleepingposition in normal infants and in infants dying suddenly and unexpectedly we conducted a case-control study of all such infant deaths in a defined part of the counties of Avon and Somerset over 18 months. Thisstudy formed part of a prospective investigation of all infant deaths in the county of Avon, which will be reported infull elsewhere. Methods We were notified of all suddenandunexpected deaths of infants (from birth to 1 year)in a defined area comprising most of the county of Avon and part of Somerset. On the day of an infant's death his or her Institute of ChildHealth, Bristol BS2 8BJ Peter J Fleming, FRCP, consultant paediatrician Yehu Azaz, MRCP, research fellow Alison Stewart, SRN, research midwife Bath Unit for Research into Paediatrics, Bath BAl 3NG Ruth Gilbert, MRCP, research fellowPeter T Rudd, MD, consultant paediatncian Bristol Maternity Hospital, Bristol BS2 8EG P Jeremy Berry, MRCPATH, consultant paediatrncpathologist Royal United Hospital, Bath BAI 3NG Elizabeth Hall, FRCPATH, consultant pathologist Correspondence to: Dr Peter Fleming, Department of ChildHealth, Bristol Maternity Hospital, Bristol BS2 8EG. BrAledi 1990;301:85-9 BMJ VOLUME 301 14 JULY 1990 85  general practitioner and health visitor were contacted. The health visitor was asked to identify from her caseload the two infants livingin the same neighbour- hood who were closestin age to the infant who had died. As part of a programme of support for bereaved families the parents of the dead baby were contacted and seen by one of us usually at home and with eitherthe general practitioner or health visitor, as soon as possible after the baby had died. We visited the parents on two to four further occasions over the next two to three months. A detailed, structured history was taken from the parents, including socialfactors, maternalmedical history, family history, and details of the pregnancy and perinatalperiod. A full medical history of the dead baby was taken with emphasis on recent signs of illness, feeding, and sleeping.Precise details were collected of the infant's last sleep, with particular note of the time and position in which the babyhad been put down, theposition in which he or she had beenfound, theprecise quantity and nature of theclothing and bedding,whether the babyhadbeen swaddled, whether the bedclothes had been over the baby s head when found, what heating was in the baby s room and the time the heating hadbeen on. Medical recordsof both the motherandbaby were used to obtain information on the pregnancy perinatal period, and any subsequentproblems, together with the infant's growth chart. As soon as possible after the death ofthe index infant we visited the two control infants at home and took an exactly comparable detailedhistory, with particular reference to the 24hours preceding the home visit. The controlinfants were weighedand microbiological samples taken for an investigation of the role of infection in the sudden infant death syndrome (to be reported in full elsewhere).   Data were obtained from the Meteorological Office on the daily maximum and minimum temperaturesrecorded in Avon over the period of the study. For each infant who had died and for each control infant these temperatures were recorded for the 24hours preceding the death or the home vi-sit respectively. All the infants who had died were discussed at a confidential meeting, at which the history and patho- logical findings (including histological, biochemical, and microbiological findings) were discussed. Each of the infants was thenassigned to one of two groups for the purpose of this study: group 1 comprised infantsfor whom a full and sufficient explanation of their death was found by pathological examinationand group 2 comprised infants for whom a full and sufficient explanation was not found, though some abnormalities that might have been contributory were identified in many infants. Group 2 thus equates to infants who had died from the sudden infant death syndrome. We tabulatedthe information on the infants' bedding and clothing and then calculated the total thermal resistance usingpublished values for the thermal resis- tance of each material.4   These values were expressed in tog units(the tog valueof a fabric is defined as 10times the temperature difference in degrees Celsius between its two faces when theheatflow is equal to 1 W/m2). We estimated the proportion of the infant'ssurface area covered by each garment or item of bedding, and these values were used to calculate an effective total thermal resistancefor the coverings on each infant. The estimatesof surface area covered by each garment were based on data producedby the International StandardsOrganisation for adult clothing modified by factors relating proportional surface areas of body parts in infants and adults. 36 Blankets and other bedding were assumed to cover 80 ofthe infant's surface area unlessthe infant was reported to havebeen completely covered. Swaddling was assumed to increasethe effective thermal resistance of the swaddlingbedding by a factor of two. The data on thermal resistance of beddingand clothing and on theproportional covering by various garments were supplied by theShirley Institute, Manchester (E Clulow, personal communication). Statistical methods used were the X2 method of discordant triplets, the Mantel-Haenszel test for comparing an index case with two matched controls, and multiple logistic regression for an index case and two controls. 7 The study was approved by thehospital ethical committees in theparticipating districts. ResultsSeventy two unexpected infant deaths that occurred from November 1987 to April 1989 were included in the study. A further six unexpected deathsoccurred of infants normally resident in the area but were not included because of the absence of one of the investi- gators (four infants) or because the death had occurred while theinfant was outside the study area (two infants). Results are given for the 72 deaths and 144 control infants. Most families of theinfants who had died were seen within24 hours of the death (50), and all were seen within72 hours. The median time from the death of the index infants until thecontrolinfants were seen was two days (92 were seen within three days and 129 within seven days). The mean ages of the index and control infants were close (94-4 and 97 0days respectively) (table I). The mean age ofthe two controlinfants was within 1 days of the age of the dead infant for 65 dead infants, and within three days for35. The minimum temperatures in the 24hours before the control infants were seen were close to those in the 24 hourspreceding the deathsof the index infants. In 48 cases the temperature difference was less than 3°C and in 54 it was less than 4°C. The median temperature difference was  C with the temperatures being lower in the periodsbefore the controls were seen. There was a weak negative relation between the minimum outside temperature in the preceding 24hours and the total thermal resistance of applied bedding and clothing for both the infants who had died and the controls, with a wide scatter at all temperatures. Table II shows the relation between social class (according to the Registrar General s classification) and total thermal resistance for bedding and clothing. Among the controls there was a slight socialclass gradient with infants insocial classes IV and   tending to havehigher values than those in social classes I and II. Thisgradient wasnot apparent among the infants who had died, for whom valuesof total thermal resistance were slightly higher for all socialclasses. From the information collected on signs of illnessin the preceding 24 hours'8 infants were identified who had shown signs of illness that have been considered to be potentially serious.'9 These signs included difficulty breathing, frequent coughing (>5 bouts in 24 hours), diarrhoea, vomiting,high temperature, irritability, lethargy, and missing more than one feed. Such signs were reported by the parents of16ofthe 72 infants who had died and 14of the 144 controls (odds ratio=2-4, TABLE II-Mean (SD) total thermal resistance of clothing and bedding (tog) for infants who had died and for control infants by social class Social class Infants who had died Control infants I and II 8-57 (3 95) (n= 15) 725 (2-77)(n=41) III 9-17 (2-7) (n=27) 8-26 (2-81) (n=66) IN' and V 8-67 (3 09) (n=9) 844 (316) (n=23) Unknown or unclassifiable 10 18 (4-45) (n=21) 75 (3-01) (n= 14) BMJ VOLUME 301 14 JULY 1990 TABLE I -Distnrbution of ages of infants who had died and control infants. Values are numbers (percentages) Infants who Control Age had died infants (weeks) (n=72) (n= 144) -4 6 (8) 10 (7) -8 18 (25) 28 (19) -12 11 (15) 24(17) -16 13 (18) 34 (24) -20 8 11 16(11) -24 5 (7) 12 (8) -28 6 (8) 8 (6)-32 5 (7) 12  8) >32 86  'FABLEiII-Position in which itnfants had been puit down to sleep Infants who Control had died infants* Position (n=67) (n= 134) Prone 62 76Side 432 Supine 1 23 Unknown 3 *X Test for discordant triplets. Relative risk 8-8 comparing pronewith other positions (95 confidence interval 7 0 to 11 0, p<OOOI). TABLE IV -Tog valzues for bedding and clothingfor infants who had died and their matched controls. Figures are numbers (percentages) Infants who Control Tog had died infants value (n=64) (n= 134)2   )2) 2 (1) 4 4 (6) 8 )6) 6 7(11) 35(26) 8 12 (19) 34(25) 10 11 (17) 22(16) 12 18(28) 21(16) 14 6  9) 8 (6)16 3  5) 3 (2)181(2) 1(1) 20 1(2) TABLE V-Tog values for bedding and clothing among infants who had slept in prone position. Figures are numbers (percentages) Infants who Control 'rog had died infants value (n=59) (n= 76) 2 1 (2) 2 (3) 4 3 (5) 5  7) 66 (10) 20 (26 8 12(20) 19(25) 10 11 (19) 16(21) 12 15 (25) 9 (12) 14 6 (10) 4  5) 16 3(5)   (1) 18 1  )2) 20   (2) 95 confidence interval 11 to5 1, X2=5S4, p=002 byMantel-Haenszel test). Valuesobtained for the total thermal resistance of beddingand clothing for the infants who had potentiallyserioussigns wereno different from thevaluesfor those without such signs foreither theinfants who had died (mean 9 3 and 9 1 tog respectively)orthecontrol infants (mean 7 8 and 8 1 tog respectively). After pathological investigation a full explanation was found for the deathsof five infants(severegastroenteritis and dehydration, septicaemia, haemorrhagic disease, perforated stomach, and Down s syndrome with atrioventricular canaldefect and cardiac failure). These infants were therefore assigned to group 1. Four of theseinfants had shown major signs of illness intheir last 24 hours. Valuesobtained for the total thermal resistance of clothing and bedding for the five infants in group 1 (median 87tog, range 5 6 to 122tog) wereno different from those for the 67 infants in group 2, whether they had signs of illness (mean 8 65, SE 0-67 tog) or not (mean 9- 1, SE 0-43 tog). For thedetailed analysis of the effects of sleeping position and total thermal resistance of bedding plus clothing only the 67 infants in group 2 (that is, those who died from the sudden infant death syndrome)and their controls were considered. Complete data on beddingand position for the index infant and the two control infants were availablefor 62 of the 67 infants who had died. Table III shows the positions in which the infants hadbeenput to sleep. Sixty two of the index infants hadbeen put to sleep prone and for 60 this was their usual sleeping position. Of 134 controls, 76 had beenput to sleep prone. This difference was highly significant  X2 test for discordant triplets, relative risk 8-8, 95 confidence interval 70 to 110, p<0 001). Table IV shows thecalculatedvalues for total thermal resistance of beddingand clothing for the infants who had died and thecontrol infants. The infants who had died were more heavily wrapped than thecontrol infants (mean thermal resistance forinfants who had died was 91v 8 0 tog for control infants). The mean difference in thermal resistance of beddingand clothing between the value for theinfants who had died and the mean value fortheir matched controls was 1-1 tog (95 confidence interval 0 15 to 2-2, p=0 025, paired t test). After allowing for thedifference in sleeping position between theinfants who had died and controls thedifference in thermal resistance of bedding plus clothing was significant by multiple logistic regression (relativerisk associated with each   tog rise above 8 tog was 1-14, 95 confidence interval1-03 to 1-28, p<0 05). A further multiple logistical regression was performed to examine theinteraction between thermal resistance of beddingand clothing and sleepingposition for three ranges of thermal resistance (<6 tog, 6-10 tog, >10 tog). This showed that the prone position and thermal resistance > 10 tog wereindependently associ- ated with an increased risk of suddenunexpected infant death, with relativerisks of 7 39 (2-57 to 21 2, p<0 001) and 7-89 (251 to 248, p<0 001) respectively when compared with supine orsideposition and thermal resistance <6 tog respectively. Thermal resistance of bedding plusclothing in the range6-10 tog was not associated withan increased risk of sudden unexpected infant death (relativerisk 1-95, 0-83 to 4.54, p>0 0S). Table V shows the total thermal resistance for bedding and clothing for those infants who had been put to bed prone. The differences between theinfants who had died and thecontrol infants increased, with a greater proportion of the dead infants havingbeen bothprone andunder excessive thermal insulation. All of the 11 infants who had died andhadhad a total thermal resistance ofclothingplus bedding greater than 12 had been prone. Of the 12 controlinfants who hadhad values greater than 12 tog, seven had been supine or on their sides  y with Yates'scorrection6-674, p<0 01). Among thecontrolinfants a higherproportion of those with more beddinghad slept supine or on theirsides thanprone (for example, 35 (19) of infants who had slept supine or on theirside hadhad bedding plusclothing with atotal thermal resistance of more than 10 tog, compared with 18 (14) of prone infants). Such a breakdown of positions was not possible for the infants who had died because few of them had been either supine or on their sides. The heating had been left on all night in the homes of a significantly higher proportion of theinfants who had died (28 of 67) than of the control infants (34 of 134)  y2 for discordant triplets, relativerisk 2 7, 95 confidence interval 1 4 to 5-2, p<0 01). The weights of the dead infants at necropsywerelowerthan those of the matched controls, though detailed assessment of the preceding growth charts did not suggestthatthere hadbeen an appreciable weight loss before death. The mean difference in weightbetween the value for the infants who had died and the mean value for the two controls was 660g (95 confidence interval 180 to 1140, p<0 01). These data will be reported in full elsewhere. In view of the difference in weightbetween the twogroups we examined the relation between the total thermal resis- tance of beddingand clothing and the infants' weight for both theinfants who had died and thecontrol infants. The controlinfants showed no significant relation between weight and total thermal resistance of beddingand clothing, but the infants who had died showed a weak but significant positive correlation (r= 0-31, p<001). Similarly,there wasno significant relation between thevalues of total thermal resistance for clothingplus bedding and age forthecontrols but there was a significant positive correlation fortheinfants who had died(r=0-42, p<0 001). Thus theolder infants who had died suddenlyandunexpectedly (who also hadbeen heavier) hadbeenmore heavily wrapped than the younger infants. Because of the bimodal distribution of theages of the infants who died (table I) theanalysisof the effects ofsleepingposition and the quantity of bedding was repeated for the younger (<70 days) and older (Be70 days) infants separately. This analysis showed thatfor the younger infantsfor whom complete data were availablethere was no significantdifference in total thermal insulation of bedding plusclothing between the infants who had died (n=24) and thecontrolinfants (x2=0-75, p>005) whereas the prone position was associated with an increased risk of suddenunexpected infant death(x2=7 14,relativerisk 4 15, 95 confidence interval 1 32 to 13 04, p<0 01). For theolder infants for whom complete data were availablethere had been a highly significant excess of bedding and clothing on theinfants who had died (n=38)compared with control infants. The mean difference in thermal resistance of bedding plus clothing between thevaluefor the infants who had died and the mean value for their controls was 2 39 tog (paired t test p<O0 1). There was a significantassociation between prone position and the risk of sudden unexpected infant death (X2= 13 9, relative risk 9 81,2 05 to 46-95, p<O OOl). A multiple logistical regression was performed on data from theseinfants toassess the interaction between sleeping position and values of thermal resistance in thethree ranges previously investigated forthe wholegroup (<6 tog, 6-10 tog, >10 tog). For the prone position compared with thesideor supine positionthe relativerisk was 15 1 (2 6 to 89 6, p<0 001). For thermal resistance of bedding plusclothing >10 tog compared withthermal resistance <6 togthe relative BMJ VOLUME 30114 JULY 1990 87  risk was 25-2 (3 7 to 169-0, p<0O001). Values of thermal resistance in the rangeof6-10 tog were not associated with an increased risk (relativerisk 4 3, 0 9 to 21-0, p>005). Discussion The relation between overheating and the sudden infant death syndrome has been suggested previously, but the evidencehas been mainly anecdotal.' Our studyhas shown that among a defined population of infantsthose who died were more heavily wrapped than control infants of the same age and in the same community seen shortly after the deathof the index infants. In addition to theexcess of clothing andbeddingon the infants who had died a higher proportion of them had been in rooms in which the heating wason all night. Wailoo et al showed a weak negative correlation between room temperature and quantity of bedding.4 Because of the natureof our study we were unable to investigate this relationdirectly but used outside temperature as an indirect measure of the likely thermal environment in the bedroom; several authors have shown a directrelation between the outside temperature and room temperature.49'0 On average the days on which the infants died were slightly warmer than the days on which thecontrols were seen, suggesting thattheir bedrooms may also havebeen warmer. Thus the effects of the relative excess ofclothing andbedding on theinfants who died were likely to have been greater. This study also confirmed the observationsof other investigators that there is an association between the prone position and the sudden infant death syndrome. In our study so few of theinfants who had died had been supine or on theirsidesthat we were unable to look indetail at the ways in which these positions interacted with other factors.20 2 In the prone position the exposed surface area of the baby that can contribute to radiantheat loss is less than that in the supine position.   In the prone position infants are also more likely to tolerate bedding rising upand covering part or all of their heads than in the supine or side position, as in these latter positions the bedding will make contact with the malar region, which is very sensitive to changes in physical contact, particularly during rapid eye movement sleep.29 For an infant who is heavily wrapped in acot, 85 of total heat loss may be through the head,46 and thus partially or completelycovering the head is likely to have a considerable effect on the infant's thermal balance and may lead to an inability to lose heat and thus to overheating. Spontaneous move- ments by infants are more likely to lead to the covers rising up over them in the prone position than in the supine or side position, in which the bedding may bethrown off by the same types of movements. We have thus confirmed the theoretical suggestion by Nelson et althat when there is excess bedding the risk of suddenunexpected death is higher forinfants who are prone.6 This risk is likely to be further increased if the metabolic rate rises, such as normally happensby 1 or 2 months of age.I Other factorsthat increase metabolic rate, such as acute viral infections, are alsolikely to increasethe risk ofoverheating. Previous studies have shown a high correlation between the presence of signs suggestive of acute viral infections and sudden infant death syndrome.'924 Inthe present study some evidence of suchan association was found,but many of thecontrolinfants had similar signs to those of theinfants who had died. Although two previous studies have shown that many parents respond to perceived illness in their infant by increasing the bedding,9 0 we foundno such trend.In New Zealand parents withhigher educational achievements wrapped their babies more than those withlower achievements, ' whereas in Exeter thereverse was true.9 Among thecontrol infants in ourstudy a trend was seenof less wrapping in the higher socioeconomic groups, as in Exeter. Among the infants who had died, however, there wasno such trend, suggesting that the subgroup of parents from socialclasses I and II who tended to overwrap theirinfants were overrepresented among the parents of infants who had died. Unexpected findings were that among theinfants who had died theolder infants tended to be more heavily wrapped than the young ones, though no such trends werenoted among the control infants, and that the increased risk of sudden unexpected death with overwrapping was significant for only the older infants. The higher ratio of mass to surface area in theolderinfants, together with their higher metabolic rates, may make themmore vulnerable to the effects ofincreased thermal insulation. There is no reason to believe that theparents of the infants who had died had increased the amounts of bedding or clothing on their infants with increasing age. Thus the same degree of overwrapping, particularly in the prone position, may be more hazardous to these slightly older infantsparticularly when they havemild viral infections, which are commonly present in older infants who die suddenly and unexpectedly. Reliable information onwhat constitutes appropriate thermal care for normal infants beyond the first month of life is scarce and many sources of information for parents and for health care professionals emphasise the risk ofcold stress but say little about overheating. For infants who sleep on their sides or supine the effects of excess bedding may be less than forinfants who sleep prone, andsome may think that this warrants wide- spread adoption of the recommendation that all infants shouldbe put to sleep supine or on their side. Infants who are supine are potentially more vulnerable to the effects of cold stress than infants who are prone,'6 and gastro-oesophageal reflux is more common and more severe in the supine position.2 In newborn preterm infants oxygenation is better in the prone position than the supine position,26 but little is known about thisin older infants. For infants with gastro-oesophageal reflux and preterm infants the prone position is preferable. For other infants the most appropriate sleepingposition may be on theirside or supine. For all infants particularattention shouldbe paid to thermal care and the avoidance of heat stress. Parents shouldbe encouraged to checkwhether their babies feel hot or cold and to adjust the bedding accordingly. This may be easier if multiple thin layers of bedding are used (for example, blankets) rather than a singlethick covering (for example, a duvet). Educating parents about sleep- ing position for and correct thermal care of their babies may help reduce the incidence of sudden infantdeaths. RG was supported by the Foundation for the Study of Infant Deaths, YA by a grant (No S/P/1705) from Action Research for the Crippled Child, and AS by a grant from Cot Death Research Appeal. We thank Maggie Shapland for help and advice on data processing andcomputing,Julie-Ann Evansand Catherine Waters for statistical advice, and Dr Jean Golding for help in planning and designing thestudy.   Bacon CJ. Overheating in infancy. Arch Dis Child 1983;58:673-4. 2 Stanton AN Scott DJ, DownhamMAPS. Is overheating a factor in someunexpected infant deaths? Lancet 1980;i: 1054-7. 3 Stanton AN. Overheatingand cot death. Lancet 1984;ii: 1199-201. 4 Wailoo MP, Petersen SA, WXhittaker H,Goodenough P. The thermal environment in which 3-4 month old infants sleep at home.Arch Dis Child 1989;64:600-4. 5 Lee NNY Chan YT, Davies DP,Lau E, Yip DCP. Sudden infant death syndrome in Hong Kong: confirmation of low incidence. Br Aled J 1989;298:721. 6 Nelson EAS, Taylor Bj, WCteatherall IL. Sleeping position and infant bedding 88 BMJ VOLUME 301 14 JULY 1990  may predispose to hyperthermiaand the sudden infant death syndrome. Lancet 1989;i:199-201. 7 Azaz Y, Fleming PJ, Levine MR McCabe R. The relationship between environmentaltemperature,metabolic rate and sleep state in infants from birth to two months [Abstract]. Early Hum Dev 1989;18:293. 8 Levine MR Fleming PJ, AzazY, McCabe R. Changes in breathing pattern accompanying environmental cooling in human infants [Abstract]. Early Hum Dez 1989;19:216. 9 Eiser C, Town C, Tripp J. Dress and care of infants in health and illness. Arch Dis Child1985;60:465-70.10 Nelson EAS, Taylor BJ.Infant clothing, bedding androom heating in an area ofhighpostneonatal mortality. Pediatric and Pernnatal Epidemiology 1989;3:146-56. 11 Gilbert R,Orrefu V, White DG Berry PJ, Fleming PJ, Rudd PJ. Detection of alpha-interferon in babies dyingsuddenly and unexpectedly. Pediatric Reviewsand Communications (inpress).12 Clulow E. Thermal insulatingproperties of fabrics.Textiles 1978;1:2. 13 International Standards Organisation Working Group. Draft report on thermalenvironments. Geneva:ISO, March 1986. (ISO/TVC/159/SC5/WGC.) 14 Boyd E. The surfacearea of the human body. Minnesota: Universityof Minnesota Press, 1935:113-42. 15 Klein AD Scammon RE. The regional growth in surfaceareaofthe human body in prenatal life. Proceedings of the Society of Experimental Biology 1930;27:463-6.16 Stothers JK,Warner RM. Thermal balance and sleep state in the newborn.Early Hum Dev 1984;9:313-22. 17 Osborn JF.Basic statistical methodsfor epidemiological studies. London: London School of Hygiene and Tropical Medicine, 1987. 18 Gilbert RE, Fleming PJ, AzazY, Rudd PT. Signs of illness preceding sudden unexpected death in infants. BrMedJ71990;300: 1237-9.19 Stanton AN Downham MAPS OakleyJR, Emery JL, Knoweldon J. Terminal symptoms in children dyingsuddenly and unexpectedly at home. BrMedJ7 1978;ii:1249-51. 20 Jonge GA, Engelberts AC, Koomen-Liefting AJM Kostense PJ. Cot death and prone sleepingposition in The Netherlands. BrMedJ 1989;298:722. 21Beal S. Sleeping position and SIDS. Lancet 1988;i:688. 22 WheldonRE. Energy balance in the newborn baby: use of a mannikin to estimate radiant and convective heat loss. Phys Med Biol 1982;27:285-96. 23 Fleming PJ, Levine MR GoncalvesA. Changes in respiratory patternresulting from the use of a facemask to record respiration in newborn infants. Pediatr Res 1982;16:1031-4.24 Knowelden J, Keeling J, Nichol JP. A multicentre study of post-neonatal mortality. London: Department of Health and Social Security, 1984. 25 Orenstein SR. Effect of non-nutritive sucking on infant gastro-esophageal reflux. Pediatr Res 1988;24:38-40. 26Orenstein SR, Whittington PF, Orenstein DM. The infant seat as treatment for gastroesophageal reflux. N EnglJ7 Med 1983;309:760-3. 27 Martin RJ, Herrell N, Rubin D, Fanaroff A. Effect of supine andprone positions on arterial oxygen tension in the preterm infant. Pediatrics 1979;63:528-31. (Accepted 20 April 1990 Sources of stressin women junior house officers Jenny Firth-Cozens Abstract Objective-To determine the causes of stressin women doctors and relate these tolevels ofdepres- sion. Design-Questionnaire study. Subjects-Of 92 women doctors who had graduated from the universities of Leeds, Man- chester, and Sheffield in 1986 andhadbeen working as junior house officers foreight months 70 (76 ) returned completed questionnaires. Main results-Mean score on the general health questionnaire was 13-79 (SD 5.20) andon the symptom checklist for depression was 1-43 (0.83). The scoresof 32 subjects (46 )were above the criterionfor clinical depression. Overwork was perceived as creating the most strain, followed by effects on personal life, serious failures of treatment, and talking to distressed relatives. Both stress and depression were related to effects on personal life, overwork, relations with consultants, and making decisions. Sex related sources of stress were con- flicts between career and personal life, sexual harass- ment at work, alack of female role models, and prejudice from patients. In addition to these, dis- crimination by senior doctors was related to depres- sion. Conclusion-Changes are needed in the career pathsof women doctors, and couldbe implemented. Introduction Symptoms of stress and depression have been found to be high in juniordoctors, both in Britain' and in North America,24 and these findings aretrue for both men and women. Several studies have shown, however, thatthe stress and depression levels of women doctors are considerablyhigher than those of other profes-sional women and of male doctors'  ; for example, in a study of junior house officers, Hsu andMarshal found that women were one and a half times more likely to be classified as depressed and eight times more likely to be severely depressed.2 In addition, women doctors in general havebeen reported to have suicide rates of up to fourtimes those of their age mates.7 Though it is always difficult to compare relatively small groupswith the general population, a recent Swedish study used a 10 year sample and comparisonswith academics to find that women doctors had higher suicide rates when com- paredwith both thegeneral population and women academics, while men doctors had rates equal to thegeneral population and higherthan thoseof academics.8 Although higher levels of occupational stress have been reported in women generally,9 a recent meta- analysis of comparisons of maleand female workers showed no differences.'0 Notman et al foundno sex differences at intake to medical school, and a British longitudinal study showed that there were no sex differences in stress or depression when the subjects were students but that higher rates ofdepression existed when they were junior house officers.' 2 Despiteany difficulties ininterpretation, it seems clearthat women doctors are an occupational group at risk for depression and suicide and itis particularly important to attend to possible reasons forthese differences.Studies comparing men and women junior doctors on work factors have found similar scores for satis- faction with career choice, perceived competence, and reported levels of fatigue,6 and no differences have been reported on job perceptions or sources of stress.  3 Studies looking specifically at women doctors, however,have reported stress arising from career and family conflict,'4'6 prejudice,5'7 and a lack of role models.   A recent British study that categorised accounts of stressful events of male and female junior house officers found not one account of suchproblems, but this may havebeen because this method reportsacute rather than chronic stressors.'3 I there- fore considered the perceived causes of stress in women doctors in more detail and relatedthese to levels ofdepression. Method A list ofhospitaladdresses of preregistration doctors who had graduated from the universities of Leeds, Manchester,and Sheffield in 1986was provided by the postgraduate offices of those universities as part of the junior doctors project.' The 92 women doctors who had not been contacted previously under that project were sentpostal questionnaires andstamped addressed envelopesalong with a letter explaining that this was a Department of Psychology,University of Leeds, Leeds Jenny Firth-Cozens, PHD, clinical psychologist BrMedj 1990;301:89-91 BMJ VOLUME 301 14 JULY1990 89
Similar documents
View more...
Search Related
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks