Quality in Health Care 1996;5:67-72

Newcastle satisfaction with nursing scales: an

instrument for quality assessments of nursing care

Lois H Thomas, Elaine McColl, Jonathan Priest, Senga Bond, Richard J Boys

AbstractObjectives-To test the validity and re-liability of scales for measuring patients'experiences of and satisfaction with nurs-ing care; to test the ability of the scales todetect differences between hospitals andwards; and to investigate whether place ofcompletion, hospital, or home influencesresponse.Design-Sample survey.Setting-20 wards in five hospitals in thenorth east ofEngland.Patients-2078 patients in general medicaland surgical wards.Main measures-Experiences of andsatisfaction with nursing care.Results-75% of patients approached tocomplete the questionnaires did so. Con-struct validity and internal consistencywere both satisfactory. Both the experi-ence and satisfaction scales were found todetect differences between randomly sel-ected wards and hospitals. A sample ofpatients (102) were sent a further ques-tionnaire to complete at home. 73%returned this; no significant differenceswere found in either experience or satis-faction scores between questionnairesgiven in hospital or at home.Conclusion-Scales to measure patients'experiences of and satisfaction withnursing in acute care have been developedand found to be valid, reliable, and able todetect differences between hospitals andwards. Questionnaires can be given beforepatients leave hospital or at home withoutaffecting scores, but those given at homehave a lower response rate.(Quality in Health Care 1996;5:67-72)

Keywords: nursing, patient satisfaction, psychometrics

IntroductionUnderstanding consumers' views is essential ifany service is to be developed or improved.The importance of understanding whatpatients think about health care is clearly ex-pressed in the United Kingdom Government'swhite paper Working for Patients,' but there isstill a gap between the objective of planningand delivering care "in a way which aims tomeet the expressed wishes of patients" and thereality of doing so.

In nursing, developments such as thenursing process2 and primary nursing3 have ascentral tenets patient participation and choicein care, and the subsequent evaluation of theeffectiveness or otherwise of care received.Patients' views of their care, summarised as

satisfaction, are the most widely used measureof patient outcome. Patient satisfaction is alsogenerally considered to be a legitimate measureof nursing quality: indeed, it has been de-scribed as the acid test which any system ofdelivering care must pass in evaluating effec-tiveness.4 The measurement of patient satis-faction forms an integral part of nursing qualityassurance initiatives.5 6However, our reviewsof studies measuring patients' satisfaction withnursing7-9 reached the conclusion that therewas no sensitive, valid, and reliable measuredeveloped from patients' perspectives ofpatient satisfaction with nursing.As a consequence, in 1993 we began a study

which aimed to develop psychometricallysound measures of patients' experiences of andsatisfaction with nursing from a patient's ratherthan a professional's perspective. Thesemeasures are intended for use by managers inquality assurance initiatives and by ward nursesand researchers to evaluate nursing inter-ventions. We developed the Newcastle satis-faction with nursing scales (NSNS), and thispaper reports their final tests.

Firstly, we sought to examine further theconstruct validity of the NSNS. Traditionally,the validity of a new measure is assessed bycomparing results with those obtained from anestablished criterion or "gold standard"measure. However, no such measure was avail-able.7-9 In such circumstances, the emphasismust be on construct validity, which can beassessed by examining expected variation indemographic and organisational variables'0;this approach is generally termed "knowngroup" validity. Previous research" '- hassuggested that older patients and patients whohave spent less time in full time education tendto express more positive views of care received.Based on evidence from the qualitative phaseof the study9 as well as previous research,8 wehypothesised that patients who could identifyone nurse responsible for their care wouldexpress more positive experiences of andgreater satisfaction with nursing care.

Secondly, we sought to show the internalconsistency of the NSNS, expressed in termsof Cronbach's at4 and correlations betweenitems and the total.'0 Internal consistency isgenerally deemed to be adequate whereCronbach's ax is greater than 0 7 and corre-lations and between items and total exceed0-4.10 Most importantly, we sought furtherevidence of the power of the NSNS to detectdifferences between wards and hospitals.A secondary aim was to assess whether the

place that the questionnaire was given affectspatients' responses. Previous research'5 has

Centre for HealthServices Research,University ofNewcastle upon Tyne,21 Claremont Place,Newcastle upon TyneLois H Thomas, researchassociateElaine McColl, seniorresearch associateJonathan Priest, juniorresearch associateSenga Bond, professor ofnursing researchDepartment ofMathematics andStatistics, UniversityofNewcastle uponTyne, Merz Court,Newcastle upon TyneRichard J Boys, lecturerin statisticsCorrespondence to:Dr L H Thomas,Centre for Health ServicesResearch, University ofNewcastle upon Tyne,21 Claremont Place,Newcastle upon TyneNE2 4AA

Accepted for publication22 January 1996

67

on Septem

ber 1, 2020 by guest. Protected by copyright.

http://qualitysafety.bmj.com

/Q

ual Health C

are: first published as 10.1136/qshc.5.2.67 on 1 June 1996. Dow

nloaded from

Thomtas, McColl, Priest, Bond, BoYs

suggested that place of questionnaire com-pletion might influence both response ratesand expressed levels of satisfaction. Forexample, patients may respond more positivelyin hospital thinking that their responses mayinfluence subsequent care. An understandingof the effects of place of giving the question-naire on response rates and response patternsis important to potential users, as there aremajor resource savings to be made if worth-while data can be collected before patientsleave hospital.

BackgroundA structured self completed questionnaire wasdeveloped by asking 150 patients, throughindividual and focus group interviews, whatthey perceived as good or bad quality nursing.9Major themes to emerge were the availabilityand attentiveness of nurses, the degree ofindividual treatment afforded to patients, theprovision of reassurance and information andthe openness or informality of nurses. Otherthemes mentioned less often included nurses'professionalism and knowledge, ward organis-ation, and environment. The NUDIST softwarepackage for qualitative data analysis'6 was usedto code and group these themes. The conceptsidentified were used to create the item pool forthe NSNS. With patients as expert informants,we ensured that the item content was import-ant and relevant to their experience of nursingin acute hospitals. The use of many represen-tative informants to generate and select itemsis a recognised means of establishing adequatecontent validity. 1'The scales went through two phases of

refinement and reduction.18 Recognised

psychometric techniques"' were used to guidethe reduction in items; those with poor internalconsistency and high non-response rates wereremoved.'8 Remaining items in each scale weresubjected separately to factor analysis, whichshowed that each was one-dimensional:patients' experiences of all aspects of nursingcare were highly correlated, as was their satis-faction with various aspects of nursing care, andindividual concepts identified in qualitativeinterviews did not emerge as separate factors.Preliminary analysis"' suggested that the scalescould detect differences in experience andsatisfaction between wards and hospitals.The scales were incorporated into a self

completed questionnaire which comprisedthree sections:

1 EXPERIENC ES OF NURSING CARE SCAI;

A series of 26 statements on aspects of nursing(outlined above) were presented and respon-dents were asked to indicate with a seven pointLikert scale (box 1), how true each was of theirown experience. To avoid affirmation bias, amixture of 15 positively and 11 negativelyworded statements were included. Responsesacross all items were summed and transformedto yield an overall experience score, with apotential range of 0- 100, where 100 representsthe best possible experience.

2 SA I ISFACTION WI I H NURSING CARE SC ALIE

Respondents rated their satisfaction withvarious aspects of nursing care, using a fivepoint Likert scale (box 2). This section com-prised 19 items. Responses across all itemswere summed and transformed to yield anoverall satisfaction score of 0-100, in which

1 Nurses gave me information just wheit I needed itDisagree Disagree a Disagree a Neither agree Agree a little Agree a lot

completely lot little nor disagree2 3 4 6

2 Nurses used to go away anidforget what patients had askedforDisagree Disagree a Disagree a Neither agree Agree a little Agree a lotcompletely lot little nor disagree

2

Agreecompletely

I

Disagreecompletely

I

3 4 5

3 Nurses checked regularly to make sure I was okayAgree a lot Agree a little Neither agree Disagree a

nor disagree little2 3 4 5

6

4 ASurses did not seem to knowz what each other zoas doingDisagree a Disagree a Neither agree Agree a little Agree a lot

lot little nor disagree2 3 4 5 6

Box I Examples front "experiences of nursing" scale

Thinking about your stay? on the ward, hozw did von feel about: Not at all Barely Quite Vern Completelvsatisfied satisfied satisfied satisfied satisfied

4 The amount nurses knew about your care 1 2 3 4 55 How quickly nurses came when you called for them 1 2 3 4 517 How willing nurses were to respond to your requests 1 2 3 4 519 Nurses' awareness of your needs 1 2 3 4 5

Box 2 Examtples front "satisfaction with nursing" scale

Agreecompletely

I

Agreecompletely

7

Disagree a Disagreelot completely6 7

Agreecompletely

7

68

I

I

on Septem

ber 1, 2020 by guest. Protected by copyright.

http://qualitysafety.bmj.com

/Q

ual Health C

are: first published as 10.1136/qshc.5.2.67 on 1 June 1996. Dow

nloaded from

Newcastle satisfaction with nursing scales: an instrument for quality assessments of nursing care

100 denoted complete satisfaction with allaspects of nursing care.

3 DEMOGRAPHIC INFORMATION

The final section elicited information about thepatient and details of the stay in hospital.

MethodsPATIENTS

Two medical and two surgical wards wereselected at random from each of five hospitalsin the north east of England (n = 20). Wedeemed that a difference of at least 5% be-tween wards in average scores on the experi-ences of nursing scale would indicate adifference in nursing care quality. Sample sizecalculations based on preliminary datasuggested an achieved sample of 80 patientsper ward was required to detect a difference ofthis magnitude with 80% power.

PATIENT RECRUITMENT

Patients were recruited on their day ofdischarge from hospital. All patients aged 18years or older who had been in the ward for twonights or more and were not too confused ortoo ill to participate were approached byindependent, trained data collectors, the studywas explained, and they were invited to takepart. Patients were encouraged to complete thequetionnaire before discharge and return it ina sealed envelope to a collection box in theward. However, we recognised that somepatients might not have enough time tocomplete the questionnaire before leaving theward. We therefore used pre-paid envelopes(addressed to the University) and 322 (24%)patients for whom these data were availablecompleted the questionnaire at home andreturned it by post. Of the total sample 558(29%) who received questionnaires in hospitaldid not fill in date of completion, so we cannotsay where they completed the questionnaire.No reminders were used for this sample.To assess whether place of completion

affected response, a subsample of respondentswere sent a second copy of the questionnaireto complete at home 10 days after discharge.According to preliminary calculations basedon data derived from previous phases ofquestionnaire development,'8 19 a sample of 70patients completing both questionnaires wasrequired to detect a difference of 5% with 80%power. Anticipating an overall response rate of70%, a random sample of 102 respondentsreceived a second, postal questionnaire; asingle reminder, with a second copy of thequestionnaire, was sent to those who had notresponded after two weeks. Retrospectivecalculations based on observed variances forthis final round of data collection indicated thatachieved sample sizes of 42 and 48 patientswould be adequate to detect an overalldifference of 5% with 80% power in experienceand satisfaction scores respectively.

ANALYSIS

Because experience and satisfaction scalescores were not normally distributed, Kruskal-Wallis one way analysis of variance was used to

find whether the scales were able to detectdifferences in patients' scores between hospi-tals and wards. To assess construct ("knowngroup") validity, patients were categorised asunder 40, 40-65, and over 65 years; wehypothesised that older patients would havehigher experience and satisfaction scores.Similarly, we hypothesized that patients whohad completed their full time education at 19years or more (suggestive of third leveleducation) would express less positive views.Finally, we expected that patients who couldidentify one specific nurse in charge of theircare would have higher experience andsatisfaction scores than those who could not orwere unsure.For the comparison between hospital and

home, sample size calculations had allowed forthe detection of a difference of 5% with 80%power. However, test-retest analysis'9 yielded95% confidence intervals (95% CIs) of -13-6to 14-0 for the experience scale and -13)8 to15*6 for the satisfaction scale. We thushypothesised that observed differences ofmorethan 14 points (on the 0-100 scale) betweenquestionnaires answered in hospital or at homewould indicate a real difference in experienceor satisfaction scores at the level of theindividual patient. The difference in scores wastherefore calculated for each respondent;because these differences had a normal distribu-tion, parametric statistical tests were used.

ResultsRESPONSE RATES

Of 2078 eligible patients, 1920 (92%) agreedto participate. An overall response rate of 75%(1559/2078) was obtained, representing 81%(1559/1920) of those agreeing to participate.Non-respondents were significantly more likelyto be women (304 (58.6%)) and were signifi-cantly older (mean age 62-2 v 59-8 for respon-dents). There was no difference in duration ofstay between the two groups. Response ratesranged from 67% to 9 1/% per ward and from69% to 89% per hospital.

CONSTRUCT VALIDITY

As there was no gold standard against which tocompare scores derived from the NSNS, weassessed construct validity by examiningexpected variation related to age, educationalattainment, and ability to identify a namednurse. Respondents' ages ranged from 18 to 97(mean age 59X8, median age 64 0) years.Nearly half the sample (707 (45X6%)) was overthe age of 65 years. Patients' age was signific-antly associated with experience scores(P < 0 00 1): older patients rated their experi-ence of nursing more positively. However,there was no association between age andsatisfaction scores (P = 0 22).A crude indicator of educational attainment

was age at completion of full time education.A small percentage of respondents (95 (6.7%))were aged 19 or over on leaving full timeeducation. This variable was found to have asignificant association with experience scores(P < 0 001) and satisfaction scores (P < 0 0 1).Those who left full time education earlier rated

69

on Septem

ber 1, 2020 by guest. Protected by copyright.

http://qualitysafety.bmj.com

/Q

ual Health C

are: first published as 10.1136/qshc.5.2.67 on 1 June 1996. Dow

nloaded from

Thomas, McColl, Priest, Bond, Boys

their experiences of and satisfaction withnursing care more positively.

Nearly half the sample (700 (47-6%)) couldidentify one particular nurse responsible fortheir care. When this was the case, patientsreported more positive experiences of nursing(P = 0 001) and rated their satisfaction with itmore highly (P < 0 001) than patients whocould not identify a particular nurse.

INTERNAL CONSISTENCY

Cronbach's ox was 0 91 for the experience scaleand 0-96 for the satisfaction scale. This indi-cates that although internal consistency wasgood, further items could be removed withoutadversely affecting the scales. However, as theobjective was to keep at least one questiontapping each concept identified by patients asimportant to the quality of nursing, no furtherquestions were removed. Correlations betweensingle items and the total ranged from 0-31 to0-69 for the experience scale (24 out of 26exceeded 0-4) and from 0 53 to 0-82 for thesatisfaction scale.

ABILITY TO DETECT DIFFERENCES

Generally patients rated their experiences ofnursing care highly at both ward (fig 1; mean84-6, median 87 8) and hospital (fig 2) level.The experience scale was found to be able todetect differences between wards (P < 0-001)and hospitals (P < 0 00 1).

Patients were also highly satisfied with theirnursing care (figs 3 and 4; mean 84- 1, median88 2). The satisfaction scale was also found todetect differences between both wards(P < 0 001) and hospitals (P < 0 00 1).To ensure that variations found between

wards were due to differences in nursing quality,

rather than in patient characteristics, analysis ofcovariance was performed with chronologicalage and age at leaving full time education ascovariates. As data were not normally distri-buted, square root transformations were per-formed on experience and satisfaction scalescores (transformation = square root of 100minus experience/satisfaction score). Differ-ences between wards and hospitals remainedhighly significant (P < 0 00 1).

PLACE WHERE QUESTIONNAIRE WAS ANSWERED

Seventy four patients (73%) returned bothquestionnaires. However, as some patients hadcompleted their first "in hospital" question-naire at home after discharge, these patientswere omitted from the analysis. The samplewas reduced to 44 patients for the experiencescale and 43 for the satisfaction scale analysis.The satisfaction scale sample size, althoughinadequate to detect differences betweenwards, was more than that required to detecta 14% difference (on the 0-100 scale) at thelevel of the individual patient, the appropriatefocus for this comparison. The relationbetween hospital and home assessments wasexamined by reference to the mean differencebetween scores in each setting for individualpatients. With a paired t test, no significantdifferences were found in either experience orsatisfaction scores between questionnairesanswered in hospital or at home (P > 0 05).

DiscussionThe final phase of our study has shown that thescales of patients' experiences of and satis-faction with nursing have good internal consis-tency and construct validity. The scales haveproved capable of detecting differences

Maximum75th percentileMedian25th percentileMinimum

ONumberHospital and wardMedian25th Percentile75th PercentileRange

81 70 37 68 90 88 81 85 90 78 66 95 79 72 80 57 87 88 91 78

11 12 13 14 21 22 23 24 31 32 33 34 41 42 43 44 51 52 53 54

87.8 90.1 91.3 90.1 84.0 86.2 85.9 89.1 91.7 89.7 89.7 87.8 85.9 85.3 82.7 84.7 89.1 88.7 84.6 89.1

77.2 79.3 78.2 77.2 70.0 80.2 78.1 79.5 83.2 83.3 80.8 78.9 74.8 73.3 73.8 74.0 82.7 80.8 75.0 76.4

94.2 96.8 96.2 96.2 92.0 92.3 91.4 94.9 96.7 95.1 96.2 95.3 94.2 92.3 90.9 90.9 96.0 95.5 91.0 94.9

54-100 62-100 57-100 57-100 40-100 66-100 62-100 63-100 64-100 67-100 60-100 57-100 48-100 54-100 53-100 58-100 66-100 60-100 57-100 53-100

Figure Box and whisker plots of experience score by hospital and ward (excluding outliers or extremes)

110 r

100 F

90K

80k-

70K

60k-

50K

a.)To

a)

a.)

U1)QLxX

40K-

30 H

20K

10

70

on Septem

ber 1, 2020 by guest. Protected by copyright.

http://qualitysafety.bmj.com

/Q

ual Health C

are: first published as 10.1136/qshc.5.2.67 on 1 June 1996. Dow

nloaded from

Newcastle satisfaction with nursing scales: an instrument for quality assessments of nursing care

110F

a)To0C,,

0)cJ

a)._a)

xwU

100

90

80

70

60

50

40

30

20

10

NumberHospital and wardMedian25th Percentile75th PercentileRange

256 344 329 288 344

1 2 3 4 5

88.5 85.9 89.7 84.6 87.8

78.7 77.6 82.1 73.9 79.5

96.2 92.5 96.2 92.3 94.2

53-100 57-100 61-100 47-100 58-100Box 3 Potential uses ofNSNS

Figure 2 Box and whisker plots of experience score by hospital

between randomly selectedhospitals. Box 3 shows potentialNSNS.As with other scales measuring

faction, our scales provide rabsolute values. It is not possible,attribute a universal or absolute iito an observed score. The scalesprovide information about nurfrom patients' perspectives. As wscores, specific items in theexample, questions regarding incan be used to monitor particul-nursing practice. Scale scores c(monitored over time to build ujnorms.20We found, as have others,"1-'

patients rated their experiencesmore positively than younger pal

patients did not, however, rate their satis-wards and faction with nursing more positively. Patientsuses of the who spent less years in formal education rated

both their experiences of and satisfaction withpatient satis- nursing more highly than those who spent-elative, not more time in full time education." Thustherefore, to sample characteristics will have a bearing oninterpretation scale scores achieved.do, however, With the advent of the Patient's Charter,"' the-sing quality concept of a named nurse should be a reality,ell as overall for every patient. We found that less than halfscale - for our sample could identify a named nurse.formation - However, those patients who could identifyar aspects of one nurse responsible for their care reportedwould also be both more positive experiences of nursing andp population greater satisfaction than those who could not.

Contributing factors may include one nurse13that older being responsible for coordination of care andof nursing carer for a specific patient and the channelling

tents. Older of information through one source.

I

v

Number 78 70 37 66 89 87 80 85 87 76 66 95 74 70 81 56 85 84 90 76

Hospital and ward 11 12 13 14 21 22 23 24 31 32 33 34 41 42 43 44 51 52 53 54

Median 94.1 94.7 94.7 89.5 79.2 82.9 82.0 89.5 93.4 89.5 93.4 93.1 88.2 88.2 81.6 85.5 93.1 87.5 84.2 89.9

25th Percentile 79.1 75.0 79.8 75.0 64.9 75.0 72.4 73.7 85.5 76.3 79.9 76.3 74.3 76.3 67.1 72.2 77.0 77.6 74.7 76.7

75th Percentile 99.0 100.0 100.0 97.4 90.8 92.2 90.8 97.4 100.0 100.0 100.0 97.4 97.6 94.7 94.7 94.7 100.0 97.4 91.8 99.7

Range 57-100 48-100 56-100 50-100 28-100 53-100 48-100 46-100 66-100 50-100 55-100 47-100 42-100 48-100 32-100 38-100 48-100 55-100 50-100 53-100

Figure 3 Box and whisker plot ofsatisfaction score by hospital and ward

* Comparing patients' experiences andsatisfaction between wards, clinicaldirectorates, and hospitals

* Longitudinal comparisons (before and after aplanned or unplanned intervention)

* As an outcome measure in randomised trialsand other effectiveness studies

* Setting ward or hospital standards (a wardcould aim for 85% of patients having a positiveexperience on questionnaire items)

* Auditing standards* Measuring hypothesised improvements innursing care following - for example, theintroduction of a "named nurse" initiative -this could be done by comparing medianexperience and satisfaction scores, but alsoindividual questions where a change isexpected.

110 F

100 K

90 H

80K

70 -

0)0)

C,,

c0.)

.4-

U)

60K

50 H

40 H

30K

20 H

10

71

on Septem

ber 1, 2020 by guest. Protected by copyright.

http://qualitysafety.bmj.com

/Q

ual Health C

are: first published as 10.1136/qshc.5.2.67 on 1 June 1996. Dow

nloaded from

Thomas, McColl, Priest, Bond, Boys

110 _

100

90 _T 800' 70Co 60

' 50

.'n 40

et) 30

20

10

0NumberHospital and wardMedian25th Percentile75th PercentileRange

251 339 323

1 2 3

93.4 82.9 92.1

77.6 72.4 80.3

100.0 93.4 100.0

44-100 41-100 51-100

acute hospital care, which can detect differ-r L ences between hospitals and wards. We have no

evidence of the transferability of the scales toother settings - for example, community care.Independent studies would be required foreach setting to ensure that scales reflect theviews of patients.The NSNS package is now available at

nominal cost. This comprises a master copy ofthe questionnaire, a user's manual, and com-puter software for data entry and analysis. Thesoftware allows the production of comparativeoutput, contrasting patients' experiences and

278 334 levels of satisfaction between clinical units.4

85.5

72.4

94.7

39-100

5

87.5

76.3

97.4

46-100

Figure 4 Box and whisker plot of satisfaction score by hospital

Place where the NSNS was answered(hospital or home) was not associated withexperience or satisfaction scores. This suggeststhat the NSNS can be distributed to patientsbefore they leave hospital without biasingeither experience or satisfaction scores. Givingthe questionnaires after discharge is more

costly, involving postal charges and managinglists of addresses. Moreover, response ratesfrom patients who have gone home are

typically lower. Our overall response rate forpatients completing questionnaires in hospitaland at home was 73%. This favourable ratecould be due to patients having had personalcontact with a researcher while still in hospital.Response rates for our test-retest analysis,'9when patients were contacted only by post,were lower (61% for the first questionnaire).We found that many patients actually

completed their "in hospital" questionnaire athome. We do not know why they did this. Itis possible that respondents were concernedabout confidentiality. Alternatively, timing ofthe distribution of the questionnaire mayinfluence when patients choose to complete it.Ideally questionnaires should be distributedsufficiently in advance of discharge to enablepatients to complete and return them while inhospital. It is not always possible in practice topredict when discharge will take place. We alsosuggest that questionnaires are distributed toan independent person rather than a memberof the ward staff, that appropriate reassurancesof confidentiality are given, and that mechan-isms for assuring confidentiality (for example,sealed envelopes) are used.We have been successful in our aim of

developing scales for measuring patients' ex-

periences of and satisfaction with nursing in

We thank all the patients who took part in the study; theDepartment of Health, and Northern and Yorkshire RegionalHealth Authority Research and Development Divisions whofunded this work; the fieldworkers who recruited patients; andCath Brennand who managed the survey and produced themanuscript. The opinions expressed are ours and do notnecessarily represent the views of the funding bodies.

1 Secretaries of State for Health. Workingfor patients. London:HMSO, 1989.

2 Yura H, Walsh MB. The nursing process. New York:Appleton-Century-Crofts, 1978.

3 Pearson A. Primary nursing. London: Croom Helm, 1988.4 Richards DA, Lambert P. The nursing process: the effect

on patients' satisfaction with nursing care. 3 Adv Nurs1987;12:559-62.

5 Lees GD, Richman J, Salauroo MA, Warden S. Qualityassurance: is it professional insurance? I Adv Nurs1987;12:719-27.

6 Moores B, Thompson AGH. What 1357 hospital inpatientsthink about aspects of their stay in British acute hospitals.JAdvNurs 1986;11:87-102.

7 Bond S, Thomas LH. Measuring patients' satisfaction withnursing care. JAdv Nurs 1992;17:52-63.

8 Thomas LH, Bond S. Outcomes of nursing care: the caseof primary nursing. Intj Nurs Stud 1991;28:291-314.

9 Thomas LH, MacMillan J, McColl E, Priest J, Hale C,Bond S. Obtaining patients' views of nursing care toinform the development of a patient satisfaction scale.International Journal Quality in Health Care 1995;7:153-63.

10 Streiner DL, Norman GR. Health measurement scales: apractical guide to their development and use. New York:Oxford University Press, 1989.

11 Greenley JR, Young TB, Schoenherr RA. Psychologicaldistress and patient satisfaction. Med Care 1982;4:373-85.

12 Cleary PD, Edgman-Levitan S, McMullen W,Delbanco TL. The relationship between reportedproblems and patient summary evaluations of hospitalcare. Qual Rev Bull 1992;February: 53-9.

13 Chang BL, Uman GC, Linn LS, Ware JEJ, Kane RL. Theeffect of systematically varying components of nursingcare on satisfaction in elderly ambulatory women. West 7Nurs Res 1984;6:367-79.

14 Cronbach LJ. Coefficient alpha and the internal structureof tests. Psychometrika 195 1;16:297-334.

15 French K. Methodological considerations in hospitalpatient opinion surveys. IntJ Nurs Stud 1981;18:7-32.

16 Richards T, Richards L, McGalliard J, et al. NUDIST 2.3user manual. Eltham: Replee P/L, 1992.

17 Wilkin D, Hallam L, Doggett A. Measures of need andoutcome for primary health care. Oxford: Oxford UniversityPress, 1992.

18 Priest J, McColl E, Thomas L, Bond S. Developing andrefining a new measurement tool. Nurs Res 1995;2:69-81.

19 Thomas L, McColl E, Boys R, Priest J, Bond S. Final reportto Northern and Yorkshire Research and DevelopmentDivision: methodological and practical validation of theNewcastle satisfaction with nursing scales. Newcastle:Centre for Health Services Research, University ofNewcastle upon Tyne, 1995.

20 Black N, Sanderson C. Day surgery: development of aquestionnaire for eliciting patients' experiences. Quality inHealth Care 1993;2:157-61.

21 Department of Health. The Patient's Charter and you.London: HMSO, 1995.

72

on Septem

ber 1, 2020 by guest. Protected by copyright.

http://qualitysafety.bmj.com

/Q

ual Health C

are: first published as 10.1136/qshc.5.2.67 on 1 June 1996. Dow

nloaded from