Validation of the Revised Stressful Life Event …...2 ComputationalandMathematicalMethodsinMedicine...

Hindawi Publishing CorporationComputational and Mathematical Methods in MedicineVolume 2013 Article ID 601640 7 pageshttpdxdoiorg1011552013601640

Research ArticleValidation of the Revised Stressful Life EventQuestionnaire Using a Hybrid Model of Genetic Algorithm andArtificial Neural Networks

Rasoul Sali1 Hamidreza Roohafza2 Masoumeh Sadeghi3 Elham Andalib4

Hassan Shavandi1 and Nizal Sarrafzadegan4

1 Industrial Engineering Department Sharif University of Technology PO Box 11365-9466 Tehran Iran2Mental Health Unit Cardiovascular Research Center Isfahan Cardiovascular Research InstituteIsfahan University of Medical Sciences PO Box 81465-1148 Isfahan Iran

3 Cardiac Rehabilitation Research Center Isfahan Cardiovascular Research InstituteIsfahan University of Medical Sciences Isfahan Iran

4Cardiovascular Research Center Isfahan Cardiovascular Research Institute Isfahan University of Medical Sciences Isfahan Iran

Correspondence should be addressed to Hamidreza Roohafza roohafzacrcmuiacir

Received 21 November 2012 Revised 2 January 2013 Accepted 2 January 2013

Academic Editor Guang Wu

Copyright copy 2013 Rasoul Sali et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Objectives Stressors have a serious role in precipitatingmental and somatic disorders and are an interesting subject formany clinicaland community-based studies Hence the proper and accurate measurement of them is very importantWe revised the stressful lifeevent (SLE) questionnaire by adding weights to the events in order to measure and determine a cut pointMethods A total of 4569adults aged between 18 and 85 years completed the SLE questionnaire and the general health questionnaire-12 (GHQ-12) A hybridmodel of genetic algorithm (GA) and artificial neural networks (ANNs) was applied to extract the relation between the stressful lifeevents (evaluated by a 6-point Likert scale) and the GHQ score as a response variable In this model GA is used in order to set someparameter of ANN for achievingmore accurate resultsResults For each stressful life event the number is defined as weight Amongall stressful life events death of parents spouse or siblings is the most important and impactful stressor in the studied populationSensitivity of 83 and specificity of 81 were obtained for the cut point 100 Conclusion The SLE-revised (SLE-R) questionnairedespite simplicity is a high-performance screening tool for investigating the stress level of life events and its management in bothcommunity and primary care settingsThe SLE-R questionnaire is user-friendly and easy to be self-administeredThis questionnaireallows the individuals to be aware of their own health status

1 Introduction

Importance of stressors and their impact on human life haveattractedmany researchersrsquo interest in the recent yearsWheninvestigating stressors their frequency and intensity are themost important characteristics that must be considered [1]Not all stressors have the same impact Some of themhave more intense impact on the individualrsquos life and somehave less The importance of the same stressor can alsobe different in various societies and cultures Evaluation ofthe importance and impact of stressors is a very attractivesubject in psychological studies and many studies have beenperformed in this field [2]

To measure stressors different scales and tools havebeen developed in developed countries [2ndash4] Recently thedeveloping and use of stress measurement tools has also beenthe subject ofmany studies in developing countries aswell [5]

The use of self-administered questionnaires is demon-strated to be a great tool to obtain useful informationabout the health status in epidemiologic studies and healthsurveys [6 7] Meaningful estimates of disease status couldbe obtained by cross-check and agreement between question-naire data and standard criterion such as medical recordsMany studies have been considered an agreement betweenquestionnaire data and a criterion standard [7ndash13] In these

2 Computational and Mathematical Methods in Medicine

studies researchers used user-friendly and easy to be self-administered questionnaires

Being user-friendly for questionnaire provides a key char-acteristic and a better chance for success in questionnaire-based researches One of the advantages of the user-friendlyquestionnaire is that individuals could become aware of theirhealth status while using them

Questionnaires should be designed in a way to be anindicator of health status or other under-research conditionsIn other words a user friendly questionnaire could easily pro-vide information about the individual health status by com-paring the scores that are calculated with its cut point

By using machine learning algorithms such as artificialneural networks (ANNs) and determining a cut point weaimed to provide a weight for stressful life events that areintroduced in the stressful life event (SLE) questionnaireThisway a revised SLE questionnaire could be used as a screeningtool to differentiate healthy individuals from those who are atrisk of a disease

2 Materials and Methods

21 Dataset The data we used in this study is a part ofthe data collected for the Isfahan Healthy Heart Program(IHHP) research IHHP is a community-based interventionprogram to prevent and control noncommunicable diseasesDetails of themethodology used in IHHP including samplingstrategies survey instruments data entry and analysis inaddition to the evaluation and followup of the subjects aredescribed by SarrafndashZadegan et al elsewhere [14 15]

In summary IHHP includes a baseline survey fourfollow-ups and eventually a final phase The baseline surveywas conducted in 2001 in 2 interventional and 1 referralregions The interventional regions include Isfahan andNajafabad We selected Arak as the referral region Fourphases of annual followup and evaluation were performedon independent samples from 2002 to 2005 The Final phasewas conducted in 2007 According to the regional populationdistribution based on the CINDI protocol a multistagecluster random sampling was applied in order to differentiateurban versus rural areas [14]

Data used for the current study include the informationfrom 4569 adults aged over 18 years who have completed thefinal phase andhave all their related data availableThe relateddata include demographic information such as age sex andeducational years We asked all participants to completethe SLE questionnaire and the general health questionnaire-12 (GHQ-12) An informed consent was obtained from allsubjects

We evaluated the frequency and intensity of the stressfullife events for participants by the SLE questionnaire whichwas developed and validated by Roohafza et al [1] Thisquestionnaire consists of eleven domains identified by fac-tor analysis including home life financial problems socialrelation personal conflict job conflict educational concernjob security loss and separation sexual life daily life andhealth concern In each domain few questions with regardto the related stressful event exist To score each stressful lifeevent we used a 6-point Likert scale (0 = never 1 = very mild

2 = mild 3 = moderate 4 = severe and 5 = very severe) ForSLE the standardized Cronbachrsquos alpha is 92

The GHQ-12 was used for the validation and evaluationof the SLE questionnaire The GHQ-12 is a well-establishedscreening tool to detect psychological morbidity in commu-nity and clinical settings such as in primary care or generalpractice [16 17] It includes 12 items that are classified in twomain areas such as the ability to carry out normal functionsand the appearance of new and distressing experience If theGHQ score is ge4 the individual has a high stress level

22 The Basic Concept of Artificial Neural Networks Inspiredfrom the biology of human brain an artificial neural network(ANN) is a parallel processing system which has an executiveperformance [18] These systems are able to learn rulesembedded in the data and due to this ability it is usuallyused as a model for complex relationships between inputsand outputs or to find patterns in data Pattern recognitionis achieved by adjusting the ANN parameters by an errorreduction method through learning ANN consists of oneinput layer one output layer and one or more hidden layersFigure 1 demonstrates an example of an ANN with an inputlayer 2 hidden layers and an output layer

In Figure 1 circles in layers show neurons The inputlayer takes input signals and transfers them to the hiddenlayers The output layer takes the outgoing signal from thelast hidden layer and presents the appropriate result

Each layer is determined by its weight matrixThematrixmultiplication operation of the various layers along withapplying transfer functions is formed network structure andmap input signals transfer to output data If the networkconsists of 119899 layers the estimated mapping function 119865 willbe

119865 (119883)

= 120595119899(119882119899119899minus1120595119899minus1(119882119899minus1119899minus2120595119899minus2(sdot sdot sdot 1205951(11986811988211sdot 119883) sdot sdot sdot)))

(1)

where119883 input signals vector 119899 total number of layers 120595119894(sdot)

transfer function layer 119894119882119894119895 weightmatrix incoming to layer

119894 from layer 119895 11986811988211 weight matrix incoming to 1st hidden

layer from input layerThe number of neurons in input and output layers

depends on the structure of the subject of interestAn overallmethod for determining the number of hidden

layers and neurons is not found in the literatures Applyingan evolutionary algorithm such the genetic algorithm (GA)and the particle swarm optimization (PSO) is one of themostwidely used methods in this field

23 The Basic Concept of Genetic Algorithm GA is a partof evolutionary computing theory that is growing rapidlyThe main idea of this algorithm lies in Darwinrsquos theory ofevolution In every step of this method an initial populationof chromosomes (initial responses) produces a new popula-tion of chromosomes (secondary responses) Repeating thisprocess and generating a new population from a previousone in each step result in population growth and optimumresponse

Computational and Mathematical Methods in Medicine 3

Output signals

Input layer 1st hidden layer 2nd hidden layer Output layer

Input signals

Figure 1 An example of an ANN with an input layer 2 hidden layers and an output layer

Calculation of fitness value

Ending condition

Selection operation

Cross-over operation

Mutation operation

Optimal solutions ReplacementYes

No

Initial population

Figure 2 Genetic algorithm cycle

The process of producing a new population from a previ-ous one uses four operators as follows Selection This oper-ator selects a number of chromosomes to produce the nextgeneration in each population The selected chromosomesare parents The probability of selection of chromosomeswith potential to reach more optimal results is higher thanothers Crossover This operator takes more than one parentsolution and produces one or more child solutions fromthem Mutation Mutation is a genetic operator that altersone or more gene values in a chromosome from its initialstate and will produce a new chromosome Mutation helps toprevent the population from stagnating at any local optimaElite It is possible that some relatively optimal chromosomesare eliminated during the selection process so by this operatora number of elite chromosomes are transferred directly to thenext population

Figure 2 demonstrates a general overview of the GAmethod

24 Problem Definition In this paper the objective is toobtain a linear relation between the stressful life events intro-duced in the SLE questionnaire and the personrsquos demographic

characteristics Demographic characteristics are consideredas inputs and the stress level as an output variable Stresslevel was measured by GHQ-12 In other words we aim todetermine the coefficients 120572

119894(119894 = 1 2 46) and 120573

119895(119895 =

1 2 3) in the following equation

119910 =12057211199091+ 12057221199092+ sdot sdot sdot + 120572

119894119909119894+ sdot sdot sdot

+ 1205724611990946+ 12057311198891+ 12057321198892+ 12057331198893

(2)

In the above equation 119909119894is intensity of 119894 the stressful life

event introduced in the SLE questionnaire If the stressorsactually exist in personrsquos life 119909

119894is set to be the mean value

of this stressful life event in the population otherwise in thecase of lack of stressor in personrsquos life 119909

119894is set to be equal to

zero Parameters 1198891 1198892 and 119889

3are representing gender age

and education level respectively Gender is a binary variableFurthermore education level is a variable that is equal toeducational years 119910 is the response variable determined byusing the following formula

119910 = GHQ score times 25 (3)


If the GHQ score is less than 4 the individual is classifiedin the low stress level group and otherwise in the high stresslevel group The cut-point score of 100 in GHO is well-definedThat means if 119910 is less than 100 the subject is healthyotherwise is at risk of disease

The purpose of this study is to measure the agreementbetween SLE questionnaire data and variable 119910 as GHQ scorechanges The SLE questionnaire data for a person is shown asfollows

(1199091 1199092 1199093 119909

119894 119909

46 1198891 1198892 1198893) (4)

Once we determined the coefficients (2) was used as ascreening tool

25 Model Development In this study we have used a hybridmodel of GA and ANN to obtain coefficients of variables(120572119894 120573119894) in (2) To extract the linear relationship between

input data and response variable ANN was used In orderto increase the efficiency of the ANN in this model someof its effective parameters such as the number of hiddenneurons are regulated by GA Use of GA methods allowsvalidation and regulation of values by searching in thedomain and for number of hidden neurons This is done in away that the relation extracted by ANN affords themaximumdiscrimination and can discriminate healthy people fromthose at risk in the best way

After determining 120572119894 in order to simplify the use of (2) as

a screening tool we changed this equation as follows

119910 = 12057411199091015840

1+ 12057421199091015840

2+ sdot sdot sdot + 120574

1198941199091015840

119894+ sdot sdot sdot + 120574

461199091015840

46

+ 12057311198891+ 12057321198892+ 12057331198893

(5)

In the above equation 120574119894is product of 120572

119894and 119894 the stressor

mean value 1199091015840119894is a binary variable that is defined as follows

1199091015840

119894= 1 if 119894th stressor exist in individual0 otherwise

(6)

We established an expert panel consisting of three psy-chiatrists two psychologists two epidemiologists and onestatistician to evaluate 120574

119894coefficients At times the expert

panel changed some coefficients (120574119894)

26 Sensitivity and Specificity Once the coefficients in theequation were defined the stress level of individuals wascalculated (5) The resulting number was compared withthe GHQ score Sensitivity and specificity are calculated bydefining the following variables (a) true positive (TP) whichmeans that individual diagnosis as sick person is truly correct(b) true negative (TN) which means that healthy individualsare correctly diagnosed as healthy (c) false positive (FP)which means that healthy individuals are incorrectly identi-fied as sick and (d) false negative (FN) whichmeans that sickindividuals are incorrectly diagnosed as healthy

In our study TP for an event is a calculated value greaterthan 100 and the GHQ score greater than 4 TN for an event isa calculated value less than 100 and theGHQscore of less than4 FN for an event is a calculated value less than 100 and the

GHQscore greater than 4 FP for an event is a calculated valueof greater than 100 and the GHQ score less than 4 Sensitivityand specificity are calculated as follows

sensitivity = Number of TPNumber of TP +Number of TN

specificity = Number of TNNumber of TN +Number of FP

(7)

3 Result

We studied 4569 individuals including a female population of2252 (492) The mean age of female and male participantswas 386plusmn 151 and 385plusmn 154 respectivelyThemean numberof educational years in the female population was 89 plusmn 48and in the male population was 72 plusmn 49 (mean plusmn standarddeviation)

The SLE questionnaire included 46 stressors which werescored by a 6-point Likert scale (0 = never 1 = very mild2 = mild 3 = moderate 4 = severe and 5 = very severe)Mean scores for stressful life events in the study populationare shown in Table 1

As shown in Table 1 some financial stressors like financialinflation and low income have the greatest mean in thestudied population The higher the average score the morethe intensity of the stressor

If zero was selected for specific stressors in the SLE ques-tionnaire the total score remains unchanged and if any of thevalues from 1 to 5 was selected the total score was replacedby the mean value of stressors In addition to other stressorsa set of vectors was considered in some cases includingdemographic characteristic and changed GHQ scoreThis setof vectors is used for training and testing neural networks

After determining 120572119894(119894 = 1 2 46) (as shown in (2))

the calculated coefficient for each stressor was multipliedby its stressor mean value At the end the resulting valuewas revised by the expert panel We sorted the final resultsfor weight of each stressor in order of their importance andimpact in Table 2

Among all stressful life events death of parents spouseor siblings was the most important and influential in ourstudied population In contrast increased working hours andconcern about addiction of a family member did not gainmuch scores and was not as important

The results for the coefficients of demographic character-istics (120573

1= 0 120573

2= 02 120573

3= 0) showed the effect of age on

stress level for individuals and number of stressors Changesin GHQ score increase by 02 per year

In this study sensitivity was 83 and specificity was 81Our result shows that the SLE-R questionnaire is a screeningtool with acceptable accuracy

4 Discussion

In this paper a revised SLE questionnaire is presented Weadded weights to the SLE questionnairersquos stressful life eventsusing a hybrid model of GA and ANN The importance ofthis promoted questionnaire is because of added weightsto stressful life events The weights calculated based on the


Table 1 Mean of stressful life events in studied population

Domain Stressful life events MeanAddiction (self or family member) 018Divorce or separation 005Concern about addiction of a family members 045

Home life Quarrels with spouse 031Being accused 029Legal problems 024Troubles with children 050Financial inflation 246Low income 206

Financial problems Get into debt 169Taking on a mortgage 096Major financial problems 108Concern about your future 134

Social relations Major social changes 077Social discrimination 115Social insecurity 101Loneliness 086Failure in achieving the life goals 104

Personal conflicts Lack of social support 079Cultural alienation 065Not having an intimate friend 050Dealing with customers 050

Job conflicts Quarrel with colleaguesboss 027Improper working place and environment 039Increased working hours 044Educational problems of children 043

Educational concerns Participation major examinations 031Failure in major examinations 018High educational expenses 050Low salary 108High responsibility job 104

Job security Concern about job future 100Job layoff 053Long-lasting unemployment 034Death of a close family member 081

Loss and separation Major disease of family members leading to hospitalization 068Death of parents spouse or siblings 021Childrenrsquos separation from family 029Birth of a child 021

Sexual life Sexual relationship problems 012Pregnancy 008Unwanted pregnancy 005

Daily life Major changes in sleeping and eating habits 071Air pollution and traffic 062

Health concerns Mild illness 100Major physical disease leading to hospitalization 032


Table 2 Stressful life event score

Stressful life events ScoreDeath of parents spouse or siblings 46Childrenrsquos separation from family 32Participation of major examinations 31Low income 30Long-lasting unemployment 29Being accused 28Educational problems of children 27Major disease of family members leading to hospitalization 27Concern about job future 26Major financial problems 24Unwanted pregnancy 22Lack of social support 21Failure in major examinations 21Failure in achieving the life goals 19Low salary 19Major physical disease leading to hospitalization 19High responsibility job 18Concern about your future 17Not having an intimate friend 17Job layoff 16Birth of a child 16Addiction (self or family member) 15Legal problems 14Sexual relationship problems 14Troubles with children 13Quarrels with spouse 12Mild illness 12Divorce or separation 11Financial inflation 11Dealing with customers 11Pregnancy 11Major social changes 10Death of close family member 10Major changes in sleeping and eating habits 10Social insecurity 9Air pollution and traffic 7Cultural alienation 6Quarrel with colleaguesboss 6Improper working place and environment 5Get into debt 4High educational expenses 4Taking on a mortgage 3Social discrimination 3Loneliness 3Concern about addiction of a family members 1Increased working hours 1

population response The SLE-R questionnaire is a well-established tool that differentiates individuals who are at riskof a health problem or disease from those who are not at risk

In this study sensitivity of 83 and specificity of 81 wereobtained for a cut-point of 100

Agreement between questionnaire data and criterionstandards is the subject of many questionnaire-based studiesThe objective of these studies is developing a diagnosis toolby determining a pattern between questionnaire data and acriterion standardMany of these studies are dealt withmulti-variables and it is highly likely that there are some interactionsbetween them ANNs are powerful tools that are used forcorrelations between known inputs and outputs and couldconsider the interaction between inputs in identified patterns

Many studies have showed high efficiency of ANNs invarious applications DeGroff et al usedANNs as a diagnostictool and thereby classified heart sound data to innocent andpathological classes [19] Sensitivity and specificity of 100have been reported in their study Wan et al introduced aprediction approach based on a questionnaire using ANNmodels for skin attribute prediction [20] The results ofapplying ANNs in the current study and other mentionedstudies show that ANN is a strong and a valuable method

The Holmes and Rahe stress scale (HRSS) is one of themost commonly and widely used quantitative measurementsof psychosocial stress It is a self-reported list of 43 commonevents associated with some degree of disruption of anindividualrsquos life HRSS assists people to discern their internalstress and understand their cumulative stress over a one-yearperiod It assigns a number to the amount of stress beingfelt by a person with no margin for differential of how aperson actually internalizes the stress Individualrsquos stress levelis determined based on the calculated value in HRSS User-friendliness is one of the prominent characteristics of HRSSthat makes it to be commonly used One of the other widelyused stress scales named Life Experiences Survey (LES) thatdeveloped by Sarason et al is a self-reported structuredinterview that assesses major life events in the past year Thisuser-friendly questionnaire includes 57 items divided intotwo sections Each participants can record positive negativeor no effect for each event on their life

The SLE-R questionnaire that is proposed in this paper isa screening tool similar to HRSS and LES Despite simplicitythese questionnaires are high-performance screening toolsfor investigating the stress levels In many cases clinicaltests are costly and time consuming and associated withuncertainty In this situation clinical screening tools can helpexperts to better identify a diagnosis Because these instru-ments are self-reported individuals could gain some infor-mation about their own health status while filling them up

In the SLE-R questionnaire a global number is deter-mined for each stressful life event and differences betweenindividuals are not consideredThis questionnaire being self-reported is another limitation for our study because theparticipant could only remember that stressor that is subjectof a question in the questionnaire The SLE-R questionnairehas not been tested for normalization The questionnaireshould be compared with other questionnaires and tested indifferent communities

In conclusion the SLE-R questionnaire is introduced asa screening tool with high sensitivity and specificity The fea-tures of this questionnaire make it a useful research tool that


could be used in clinical and primary care settings Offeringclinical diagnostics to communities is very costly and notpractical Screening tools such as the SLE-R questionnaireare required to consider a smaller target group The SLE-R questionnaire can be completed by individuals with lowliteracy

Acknowledgments

This study was supported by the Isfahan CardiovascularResearch Institute The authers would like to thank theirexpert panel for the invaluable cooperation Ali Abbasaliz-adeh Abbas Attari Hamid Afshar and Mahjoobeh Kaviani

References

[1] H Roohafza M Ramezani M Sadeghi M Shahnam BZolfagari and N Sarafzadegan ldquoDevelopment and validationof the stressful life event questionnairerdquo International Journal ofPublic Health vol 56 no 4 pp 441ndash448 2011

[2] T H Holmes and R H Rahe ldquoThe social readjustment ratingscalerdquo Journal of Psychosomatic Research vol 11 no 2 pp 213ndash218 1967

[3] B S Dohrenwend L Krasnoff A R Askenasy and B PDohrenwend ldquoExemplification of a method for scaling lifeevents the Peri life events scalerdquo Journal of Health and SocialBehavior vol 19 no 2 pp 205ndash229 1978

[4] I G Sarason J H Johnson and J M Siegel ldquoAssessing theimpact of life changes development of the life experiencessurveyrdquo Journal of Consulting and Clinical Psychology vol 46no 5 pp 932ndash946 1978

[5] T Almadi I Cathers AMHamdanMansour andCMChowldquoAn Arabic version of the perceived stress scale translation andvalidation studyrdquo International Journal of Nursing Studies vol49 no 1 pp 84ndash89 2012

[6] T L Bush S RMiller A LGolden andWEHale ldquoSelf-reportand medical record report agreement of selected medical con-ditions in the elderlyrdquoAmerican Journal of Public Health vol 79no 11 pp 1554ndash1556 1989

[7] N Haapanen S Miilunpalo M Pasanen P Oja and I VuorildquoAgreement between questionnaire data and medical records ofchronic diseases in middle-aged and elderly Finnish men andwomenrdquo American Journal of Epidemiology vol 145 no 8 pp762ndash769 1997

[8] K C Barnes L R Freidhoff E M Horowitz et al ldquoPhysician-derived asthma diagnoses made on the basis of questionnairedata are in good agreement with interview-based diagnoses andare not affected by objective testsrdquo Journal of Allergy andClinicalImmunology vol 104 no 4 pp 791ndash796 1999

[9] S N BaxiW J Sheehan J M Gaffin et al ldquoAgreement betweenparent and student responses to an asthma and allergy ques-tionnaire in a diverse inner-city elementary school populationrdquoAnnals of Allergy Asthmaamp Immunology vol 107 no 4 pp 371ndash373 2011

[10] L Matthews C Hankey V Penpraze et al ldquoAgreement ofaccelerometer and a physical activity questionnaire in adultswith intellectual disabilitiesrdquo Preventive Medicine vol 52 no 5pp 361ndash364 2011

[11] S S Merkin K Cavanaugh J C Longenecker N E Fink A SLevey and N R Powe ldquoAgreement of self-reported comorbid

conditions withmedical and physician reports varied by diseaseamong end-stage renal disease patientsrdquo Journal of ClinicalEpidemiology vol 60 no 6 pp 634ndash642 2007

[12] Y Okura L H Urban D W Mahoney S J Jacobsen and R JRodeheffer ldquoAgreement between self-report questionnaires andmedical record data was substantial for diabetes hypertensionmyocardial infarction and stroke but not for heart failurerdquoJournal of Clinical Epidemiology vol 57 no 10 pp 1096ndash11032004

[13] L Hedman A Bjerg M Perzanowski and E Ronmark ldquoGoodagreement between parental and self-completed questionnairesabout allergic diseases and environmental factors in teenagersrdquoJournal of Clinical Epidemiology vol 63 no 7 pp 783ndash789 2010

[14] N Sarraf-Zadegan G Sadri H Malek Afzali et al ldquoIsfa-han Healthy Heart Programme a comprehensive integratedcommunity-based programme for cardiovascular disease pre-vention and control Design methods and initial experiencerdquoActa Cardiologica vol 58 no 4 pp 309ndash320 2003

[15] N Sarrafzadegan A Baghaei G Sadri et al ldquoIsfahan healthyheart program evaluation of comprehensive community-based interventions for non-communicable disease preven-tionrdquo Prevention and Control vol 2 no 2 pp 73ndash84 2006

[16] D P Goldberg and V F Hillier ldquoA scaled version of the GeneralHealth Questionnairerdquo Psychological Medicine vol 9 no 1 pp139ndash145 1979

[17] A Montazeri A M Harirchi M Shariati G Garmaroudi MEbadi and A Fateh ldquoThe 12-item General Health Question-naire (GHQ-12) translation and validation study of the Iranianversionrdquo Health and Quality of Life Outcomes vol 1 article 662003

[18] S Haykin Neural Networks A Comprehensive foundationMacMillan New York NY USA 1994

[19] C G DeGroff S Bhatikar J Hertzberg R Shandas L Valdes-Cruz and R L Mahajan ldquoArtificial neural networkmdashbasedmethod of screening heart murmurs in childrenrdquo Circulationvol 103 no 22 pp 2711ndash2716 2001

[20] W Wan H Xu W Zhang X Hu and G Deng ldquoQuestion-naires-based skin attribute prediction using Elman neural net-workrdquo Neurocomputing vol 74 no 17 pp 2834ndash2841 2011

Submit your manuscripts athttpwwwhindawicom

Stem CellsInternational

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014


MEDIATORSINFLAMMATION

of


Behavioural Neurology

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Immunology ResearchHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary and Alternative Medicine

Volume 2014Hindawi Publishing Corporationhttpwwwhindawicom


studies researchers used user-friendly and easy to be self-administered questionnaires

Being user-friendly for questionnaire provides a key char-acteristic and a better chance for success in questionnaire-based researches One of the advantages of the user-friendlyquestionnaire is that individuals could become aware of theirhealth status while using them

Questionnaires should be designed in a way to be anindicator of health status or other under-research conditionsIn other words a user friendly questionnaire could easily pro-vide information about the individual health status by com-paring the scores that are calculated with its cut point

By using machine learning algorithms such as artificialneural networks (ANNs) and determining a cut point weaimed to provide a weight for stressful life events that areintroduced in the stressful life event (SLE) questionnaireThisway a revised SLE questionnaire could be used as a screeningtool to differentiate healthy individuals from those who are atrisk of a disease

2 Materials and Methods

21 Dataset The data we used in this study is a part ofthe data collected for the Isfahan Healthy Heart Program(IHHP) research IHHP is a community-based interventionprogram to prevent and control noncommunicable diseasesDetails of themethodology used in IHHP including samplingstrategies survey instruments data entry and analysis inaddition to the evaluation and followup of the subjects aredescribed by SarrafndashZadegan et al elsewhere [14 15]

In summary IHHP includes a baseline survey fourfollow-ups and eventually a final phase The baseline surveywas conducted in 2001 in 2 interventional and 1 referralregions The interventional regions include Isfahan andNajafabad We selected Arak as the referral region Fourphases of annual followup and evaluation were performedon independent samples from 2002 to 2005 The Final phasewas conducted in 2007 According to the regional populationdistribution based on the CINDI protocol a multistagecluster random sampling was applied in order to differentiateurban versus rural areas [14]

Data used for the current study include the informationfrom 4569 adults aged over 18 years who have completed thefinal phase andhave all their related data availableThe relateddata include demographic information such as age sex andeducational years We asked all participants to completethe SLE questionnaire and the general health questionnaire-12 (GHQ-12) An informed consent was obtained from allsubjects

We evaluated the frequency and intensity of the stressfullife events for participants by the SLE questionnaire whichwas developed and validated by Roohafza et al [1] Thisquestionnaire consists of eleven domains identified by fac-tor analysis including home life financial problems socialrelation personal conflict job conflict educational concernjob security loss and separation sexual life daily life andhealth concern In each domain few questions with regardto the related stressful event exist To score each stressful lifeevent we used a 6-point Likert scale (0 = never 1 = very mild

2 = mild 3 = moderate 4 = severe and 5 = very severe) ForSLE the standardized Cronbachrsquos alpha is 92

The GHQ-12 was used for the validation and evaluationof the SLE questionnaire The GHQ-12 is a well-establishedscreening tool to detect psychological morbidity in commu-nity and clinical settings such as in primary care or generalpractice [16 17] It includes 12 items that are classified in twomain areas such as the ability to carry out normal functionsand the appearance of new and distressing experience If theGHQ score is ge4 the individual has a high stress level

22 The Basic Concept of Artificial Neural Networks Inspiredfrom the biology of human brain an artificial neural network(ANN) is a parallel processing system which has an executiveperformance [18] These systems are able to learn rulesembedded in the data and due to this ability it is usuallyused as a model for complex relationships between inputsand outputs or to find patterns in data Pattern recognitionis achieved by adjusting the ANN parameters by an errorreduction method through learning ANN consists of oneinput layer one output layer and one or more hidden layersFigure 1 demonstrates an example of an ANN with an inputlayer 2 hidden layers and an output layer

In Figure 1 circles in layers show neurons The inputlayer takes input signals and transfers them to the hiddenlayers The output layer takes the outgoing signal from thelast hidden layer and presents the appropriate result

Each layer is determined by its weight matrixThematrixmultiplication operation of the various layers along withapplying transfer functions is formed network structure andmap input signals transfer to output data If the networkconsists of 119899 layers the estimated mapping function 119865 willbe

119865 (119883)

= 120595119899(119882119899119899minus1120595119899minus1(119882119899minus1119899minus2120595119899minus2(sdot sdot sdot 1205951(11986811988211sdot 119883) sdot sdot sdot)))

(1)

where119883 input signals vector 119899 total number of layers 120595119894(sdot)

transfer function layer 119894119882119894119895 weightmatrix incoming to layer

119894 from layer 119895 11986811988211 weight matrix incoming to 1st hidden

layer from input layerThe number of neurons in input and output layers

depends on the structure of the subject of interestAn overallmethod for determining the number of hidden

layers and neurons is not found in the literatures Applyingan evolutionary algorithm such the genetic algorithm (GA)and the particle swarm optimization (PSO) is one of themostwidely used methods in this field

23 The Basic Concept of Genetic Algorithm GA is a partof evolutionary computing theory that is growing rapidlyThe main idea of this algorithm lies in Darwinrsquos theory ofevolution In every step of this method an initial populationof chromosomes (initial responses) produces a new popula-tion of chromosomes (secondary responses) Repeating thisprocess and generating a new population from a previousone in each step result in population growth and optimumresponse


Output signals


Input signals



Ending condition

Selection operation


Mutation operation


No

Initial population






119894(119894 = 1 2 46) and 120573

119895(119895 =


119910 =12057211199091+ 12057221199092+ sdot sdot sdot + 120572

119894119909119894+ sdot sdot sdot

+ 1205724611990946+ 12057311198891+ 12057321198892+ 12057331198893

(2)













(1199091 1199092 1199093 119909

119894 119909

46 1198891 1198892 1198893) (4)






119910 = 12057411199091015840

1+ 12057421199091015840


1198941199091015840

119894+ sdot sdot sdot + 120574

461199091015840

46

+ 12057311198891+ 12057321198892+ 12057331198893

(5)




1199091015840


(6)









(7)

3 Result









1= 0 120573

2= 02 120573




4 Discussion





























Acknowledgments


References



























of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of

















Output signals


Input signals



Ending condition

Selection operation


Mutation operation


No

Initial population






119894(119894 = 1 2 46) and 120573

119895(119895 =


119910 =12057211199091+ 12057221199092+ sdot sdot sdot + 120572

119894119909119894+ sdot sdot sdot

+ 1205724611990946+ 12057311198891+ 12057321198892+ 12057331198893

(2)













(1199091 1199092 1199093 119909

119894 119909

46 1198891 1198892 1198893) (4)






119910 = 12057411199091015840

1+ 12057421199091015840


1198941199091015840

119894+ sdot sdot sdot + 120574

461199091015840

46

+ 12057311198891+ 12057321198892+ 12057331198893

(5)




1199091015840


(6)









(7)

3 Result









1= 0 120573

2= 02 120573




4 Discussion





























Acknowledgments


References



























of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of



















(1199091 1199092 1199093 119909

119894 119909

46 1198891 1198892 1198893) (4)






119910 = 12057411199091015840

1+ 12057421199091015840


1198941199091015840

119894+ sdot sdot sdot + 120574

461199091015840

46

+ 12057311198891+ 12057321198892+ 12057331198893

(5)




1199091015840


(6)









(7)

3 Result









1= 0 120573

2= 02 120573




4 Discussion





























Acknowledgments


References



























of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of











































Acknowledgments


References



























of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of





















of






Disease Markers



OncologyJournal of





PPAR Research



Journal of

ObesityJournal of
















Validation of the Revised Stressful Life Event …...2 ComputationalandMathematicalMethodsinMedicine...

Documents

Transcript of Validation of the Revised Stressful Life Event …...2 ComputationalandMathematicalMethodsinMedicine...