Information needs and prostate cancer: the development of a systematic means of identification

Authors


Jeremy Dale, Centre for Primary Health Care Studies, University of Warwick, Coventry, CV4 7AL, UK.
e-mail: Jeremy.Dale@warwick.ac.uk

Abstract

OBJECTIVE

To design, from first principles, a valid and reliable scale for assessing the importance of specific items of information needed by patients with prostate cancer that would be straightforward to use in clinical settings, as despite its prevalence, there is little research focusing specifically on the information needs associated with prostate cancer.

PATIENTS AND METHODS

Several stages of consultation and modification were used to inform the development of a scale which was then piloted on 96 patients with prostate cancer. Respondents were asked to rate the importance they placed on a range of prostate cancer-related topics of information, and the extent to which they felt these information needs had been met. The construct and content validity of the instrument were established and an exploratory factor analysis used to guide restructuring of the tool. Internal consistency/reliability was calculated using Cronbach's α.

RESULTS

Using the scale showed that men with prostate cancer placed considerable importance on a broad range of information needs, most of which had been inadequately met. Age had a significant influence on the overall importance attributed to information, with younger patients having more need for information. The construct and content validity of the instrument were established. The factor analysis revealed four discrete factors which together explained > 68% of the variance, termed ‘basics of prostate cancer care’, ‘disease management’, ‘physical well-being’ and ‘self-help’. Internal consistency/reliability was satisfactory (α = 0.91).

CONCLUSIONS

The basis of a tool capable of ascertaining the information needs of patients with prostate cancer was developed; it may offer clinicians a valid means of ascertaining information preferences and hence potentially enhance the quality of service provided. Further research is now required to refine the tool and test the effect of its longitudinal use in clinical practice on patient satisfaction and outcome.

INTRODUCTION

Prostate cancer presents a growing health problem as longevity increases; it is now the most commonly diagnosed cancer in both the USA and the UK, and the second leading cause of cancer-related deaths in men in both countries [1]. Until recently it was a disease of elderly men, frequently associated with the development of bony metastases and typically proceeding insidiously. Men often presented late with disease that tended to be incurable [2,3]. The advent of cancer-awareness programmes and improved approaches to screening means that patients are increasingly diagnosed early, and so are living longer with their prostate cancer [4].

It is recognized that patients with cancer who are better informed tend to have better psychosocial health outcomes [5–7] and are better equipped to cope [8–10]. However, patients often fail to satisfy their information needs [11,12]. There is evidence that patients prefer to receive as much information as possible about their condition, and generally desire more information than they have actually received [8,9,13,14].

Despite the prevalence of prostate cancer, there is a lack of research focusing specifically on the information needs associated with this disease. Many patients and their families are profoundly affected by the diagnosis of prostate cancer, and once the initial state of shock at diagnosis has subsided patients begin to seek information about their disease and the therapeutic options [15]. Competing treatment options and uncertainty about the expected course of the disease may provoke depressive symptoms and overwhelming anxiety [16].

Patients are often faced with difficult decisions about treatment because of the unconfirmed benefits and side-effects of different options [17]. An understanding of the impact on quality of life produced by various treatments is necessary for informed involvement in decision-making [18], but fear and apprehension associated with cancer often hinder individuals from asking for specific information [19].

Three Canadian studies assessed the overall information needs of patients with prostate cancer. One study found that patients preferred receiving detailed information at an early stage in their care; their age and their treatment did not affect their preferences [20]. Another found that most men wanted information at all stages of illness, focusing on treatment, survival, self-care and empowerment [21]. Men receiving information tailored to their needs assumed a significantly more active role in treatment decision-making, and had lower anxiety levels [22]. It is important to accommodate information needs and decision preferences of individual patients with prostate cancer at different stages of their illness [21].

A greater understanding of the information needs of patients should make the process of providing information more worthwhile and effective for both patients and healthcare professionals, and may lead to more effective care and efficient use of resources [23]. The aim of the present study was to design a valid and reliable means of assessing the importance of specific items of information that would be straightforward to use in clinical settings; such a tool has not been available before.

Templeton and Coates [24] described the ‘adaptation of an instrument to measure the informational needs of men with prostate cancer’, a modification of the breast cancer version of the Toronto Informational Needs Questionnaire. This instrument was developed to assess the information needs of those on hormonal replacement therapy, but it has several omissions, such as in relation to psychosocial and sexual issues, that limit its usefulness.

As Duffy and Fallowfield [25] stress, tools for ascertaining patient needs and concerns need to be cancer-specific, and culturally sensitive [26]. Hence, in the present study we sought to develop a tool from first principles, involving patients with prostate cancer at the earliest stages, so avoiding the limitations intrinsic to previous attempts to apply assessment tools that have been developed for other conditions or populations. We report here the initial development of such a tool which, with further piloting, could provide healthcare professionals with a straightforward means of identifying patients’ information needs.

PATIENTS AND METHODS

The design of the scale involved several stages, from reviewing relevant research publications related to cancer in general and prostate cancer specifically, followed by open interviews with patients and healthcare professionals, leading to more formal focus groups with patients, and to the eventual design and piloting of an assessment tool.

To generate a conceptual framework of the information needs of patients, W.J. attended four hospital urology outpatient clinics in the West Midlands, observed consultations, talked to patients informally about their care and information needs, and discussed individual cases with hospital doctors and nursing staff. Issues were identified and, together with those identified from the literature review, used to create an initial list of information-related topics that could be tested in focus group discussions.

FOCUS GROUPS

Four focus groups were held in a hospital setting based on this agenda. Discussions concerned patients’ experiences of their care, and especially their information needs. In all, 26 patients attending urology clinics for prostate cancer follow-up were systematically approached and invited to participate, and 21 agreed to do so. Each focus group lasted 60–90 min and generated lively discussions among the participants. They were tape-recorded and transcribed verbatim, and the transcripts analysed using qualitative content analysis to identify important issues and recurrent themes relating to information needs. An indexing system was used to categorize data, and this process continued until consistent patterns were developed, which were then used to determine the major themes of the data. The themes that emerged were: (i) General information needs for patients with prostate cancer, including prostate function, and causes, types and stages of prostate cancer; (ii) disease monitoring and screening; (iii) treatment options, their benefits and complications; (iv) quality of life, sexuality and problems with self-management; (v) clinical deterioration and emergencies; (vi) access to information; (vii) healthy lifestyle, nutrition and complementary medicine.

SCALE PROTOTYPE

A draft questionnaire was constructed to test the importance individuals based on the emergent themes. It was first presented to urology department doctors and nursing staff for feedback and comments, and then piloted on 12 consecutive patients attending the clinic as part of a direct interview after their hospital appointment. The interview explored the scale's general content, specific items, format, wording, sequence and instructions. This provided an opportunity to ascertain the ‘face validity’ of the tool [27].

The pilot showed that the scale could easily be completed during the waiting time in the outpatient department, but that some items were felt to be unclear or superfluous. As a result, these were made more explicit or eliminated. The final version of the scale comprised 20 items (Table 1).

Table 1.  The rotated factor-loading matrix
QuestionComponent
1234
  1. Extraction method: principal component analysis; Rotation method: varimax with Kaiser normalization. The rotation converged in six iterations.

 1. Prostate cancer in general  0.7010.4190.0270.034
 2. Information/education for family/spouse  0.6010.3100.3510.165
 3. The natural function of prostate cancer  0.7640.0140.0620.159
 4. What prostate cancer is  0.7530.1750.3680.080
 5. The causes of prostate cancer  0.5310.3440.4260.047
 6. The stages of prostate cancer  0.7060.2210.2940.123
 7. Screening methods for prostate cancer  0.7200.1410.0060.169
 8. Monitoring of prostate cancer  0.5260.3560.292−0.042
 9. Treatment choices  0.3290.6990.0280.014
10. How treatment for prostate cancer works  0.3410.5940.3310.092
11. Alternative and complementary therapies  0.3080.2920.1070.616
12. How prostate cancer affects quality of life  0.2450.4090.6530.025
13. How prostate cancer affects sex life  0.2220.0500.7220.115
14. How to manage problems with prostate cancer in daily life  0.1190.4620.5820.335
15. Warning signs when prostate cancer gets worse  0.2060.6610.2700.175
16. Emergencies that can arise from prostate cancer−0.0170.6220.2840.449
17. What to do in an emergency  0.1870.7080.0260.182
18. How to access information on prostate cancer  0.5060.1940.1000.461
19. Active lifestyle  0.0800.0230.6540.494
20. Healthy food and nutrition  0.0910.0990.1730.858

The complete scale contained the following elements: (i) the information sheet and consent form, explaining the purpose of the study and that responses are confidential and will not influence the patient's care; (ii) a set of 20 items related to specific items of information, loosely ordered to reflect the major themes identified from the focus group study, each with four-point Likert scales, i.e. the importance that the subject placed on receiving that item of information (1 ‘very important’, 2 ‘important’, 3 ‘not so important’, 4 ‘completely unimportant’) and the extent to which they felt that they had received sufficient information (1 ‘too much information’, 2 ‘enough information’, 3 ‘not enough information’, 4 ‘no information at all’). Four-point Likert scales were chosen to ensure that respondents could not take a ‘neutral’ position, so forcing assignation of the importance of information items as either important or unimportant; (iii) demographic data, including age, date of diagnosis, social circumstances, previous formal education, and treatment type; (iv) the Hospital Anxiety and Depression Scale to assess levels of depression and anxiety [28].

STUDY SETTING

All patients who met the inclusion criteria (see below) at sampled urology outpatient clinics were approached while they waited for their appointment and invited to participate in the study. Data were collected over a 5-month period between April and August 2001, allowing about half of all those on the hospital's prostate cancer register to be eligible for inclusion.

Patients included in the study were those who met the following criteria. Inclusion; a diagnosis of prostate cancer based on histology or biochemistry (substantial PSA increase) and radiology; attending for prostate cancer follow-up during the data collection period; and written consent. The exclusion criteria were newly diagnosed patients, or with mental disability, or with no written consent. It was felt that it would be inappropriate to invite newly diagnosed patients to participate in this research study and an arbitrary threshold of 4 weeks since diagnosis was used.

STATISTICS

The data collected were analysed using a computer software statistics package. Because there is no reference standard against which to test the scale, it was not possible to assess criterion validity. Hence, evidence of construct validity was sought through factor analysis. A factor was considered to be important and its items worthy of retention in the scale if its eigenvalue was > 1.0 [29]. Reliability was tested using Cronbach's α to produce an estimate of internal consistency based on all possible correlations between all the items within a scale [29]. The threshold for the minimally acceptable level of internal consistency reliability was taken as 0.70 [30]. Responses were then coded to allow numerical analysis, as described above; the mean scores for importance and amount of information received for each item were calculated.

Patient attributes including age, education, duration of illness, type of treatment, and anxiety and depression scores were used to identify different subgroups of the sample, that were then compared according to the mean importance attributed to the various topics of information. A ‘discrepancy score’ was created for items where there was a mismatch between the attributed level of importance and the reported information received. A discrepancy was defined as occurring if an item was ‘very important’ or ‘important’, but reported as having received ‘not enough information’. Stepwise multiple regression analysis was used to investigate which factors significantly affected the importance attributed to each component. Factors with the a significance level of < 0.10 were excluded.

RESULTS

In all, 17 urology outpatient clinics were sampled, giving 107 patients who were invited to participate; of these, 96 (90%) consented. Reasons given for not participating were: four patients (3.7%) said they were not interested and gave no reasons; two (1.9%) had no reading glasses, did not want to post the scale and refused help with its completion; two (1.9%) had severe arthritis of their hands, and refused help with the completion; three (2.8%) said they lacked the time to participate.

The mean (range) age of the study participants was 73 (57–93) years; most (87%) had completed formal education by the age of 17 years. A few had a higher qualification, a first degree or a higher degree. The mean duration since diagnosis of prostate cancer was 32 (1–103) months.

Most patients (66%) were currently receiving some form of palliative treatment (63% hormone injections and 3% orchidectomy). A further 22% of patients had received curative treatment with no relapse (13% radical radiotherapy, 9% prostatectomy). Eleven men (12%) were under surveillance by watchful waiting and therefore received no treatment.

INFORMATION NEEDS

As Table 2 shows, all items were rated as being important/very important by ≥ 70% of respondents, with 16 of 20 items rated as being very important by at least half the participants. The greatest importance was placed on information relating to aspects of disease management, especially about the recognition and treatment of emergencies related to prostate cancer. Lower levels of importance were placed on aspects of physical well-being (e.g. how prostate cancer affects sex life or active lifestyle) or self-help (e.g. the role of alternative and complementary therapies, or healthy food and nutrition).

Table 2.  Comparison of the importance men with prostate cancer place on different items of information, and the extent to which they feel such information has been provided
 Degree of importance/amount of information
Very/too muchYes/enoughNot very/not enoughNot at all/none
Warning signs78/114/19 1/49 0/21
What to do in an emergency71/121/16 1/43 0/30
Emergencies that can arise66/128/15 0/49 0/26
Prostate cancer in general63/026/43 4/35 1/9
Treatment choices61/028/45 2/38 0/4
Causes61/021/20 8/37 1/22
Monitoring59/124/52 5/23 0/6
Screening methods57/025/40 6/29 2/12
How treatment works56/030/41 4/37 1/7
Quality of life54/130/32 6/38 1/12
What the prostate is50/032/37 4/34 1/9
Problems in daily life49/137/28 6/42 1/12
Healthy food and nutrition48/037/36 7/35 1/11
Stages48/138/23 3/46 1/12
Accessing information47/037/23 8/40 0/19
For family/spouse46/035/38 6/33 5/9
Active lifestyle42/142/34 6/30 3/15
Natural function34/142/3511/22 3/16
Sex life33/027/2622/21 11/11
Alternative & complimentary therapies32/036/1512/38 5/14

Overall, respondents felt that insufficient information had been provided for almost all items, while almost none felt they had received too much information (Table 2). There was only one item (‘monitoring of prostate cancer’) for which most patients felt that they had received enough information.

Table 3 shows the mean importance scores given to each item, and the number of respondents with a discrepancy between the importance placed on the item and the amount of information they reported having received. The items rated as having greatest importance (‘signs of worsening prostate cancer’, ‘what to do in an emergency’ and ‘emergencies that can arise’) were also those that scored highest for discrepancies. About 80% of respondents who saw these items as ‘important’ or ‘very important’ described having received insufficient information.

Table 3.  The mean importance score, the number of respondents rating the item as important and the number (%) of these who had reported having received a lack of information (discrepancy score) for each item in scale
ItemMean importance scoreN respondents rating item as importantN (%) respondents rating item as important + reporting having insufficient information
Emergencies that can arise from prostate cancer1.309075 (83.3)
What to do in an emergency1.258973 (82.0)
Warning signs when prostate cancer gets worse1.178970 (78.7)
The causes of prostate cancer1.437959 (74.6)
How to access information on prostate cancer1.588259 (72.0)
The stages of prostate cancer1.518158 (71.6)
How to manage problems with prostate cancer in daily life1.558254 (65.8)
Alternative and complementary therapies1.826752 (77.6)
How prostate cancer affects quality of life1.488250 (60.9)
Healthy food and nutrition1.578246 (56.1)
Active lifestyle1.657945 (57.0)
Prostate cancer in general1.388744 (50.5)
How treatment for prostate cancer works1.448544 (51.7)
What prostate cancer is1.488043 (53.8)
For family/spouse1.628042 (52.6)
Treatment choices1.358742 (48.3)
Screening methods for prostate cancer1.468141 (50.6)
The natural function of prostate cancer1.787338 (52.0)
How prostate cancer affects sex life2.005832 (55.2)
Monitoring of prostate cancer1.398129 (35.8)

FACTOR ANALYSIS OF THE ITEMS EVALUATING IMPORTANCE OF INFORMATION

A factor analysis of the responses given to the 20 information items within the scale showed strong correlations (coefficient > 0.25) between items 1–8 and 18, items 9, 10, 15, 16, 17, items 12, 13, 14, 19 and items 11 and 20. Principal component analysis was then used, with a factor considered as important and its items worthy of retention if its eigenvalue was > 1.0. Four factors were extracted from the 20 items of the scale, with eigenvalues of 1.06, 1.27, 1.80 and 8.26. The four factors together explained over 68% of the variance. The fifth factor produced an eigenvalue of 0.99 and hence was just below the borderline for inclusion. However, it added little to either the amount of variance explained by the model or to the clarity of the conceptual framework created and hence was excluded.

A factor analysis using an orthogonal rotation to derive the factor-loading matrix, used to reduce the within-subject correlation and to help overcome the problem of some variables loading onto two or more factors, is shown in Table 1 as the rotated factor-loading matrix. By using a threshold of 0.5, four principal factors were identified, and termed: Factor 1, basics of prostate cancer care (nine items: nos 1–8 and 18); Factor 2, disease management (five items, nos 9, 10, 15–17); Factor 3, psychosocial and physical well being (four items, 12–14, and 19); and Factor 4, self-help (two items, 11 and 20).

Table 4 compares the mean importance score for each of the four factors in relation to various variables. ‘Disease management’ was the factor with greatest mean importance score, followed by ‘basics of prostate cancer care’. Age had a significant influence on the overall importance attributed to information, with younger patients having greater information needs. Similarly, those receiving curative treatments also placed higher importance on information topics. There was also a trend towards those who were not depressed to place greater importance on topics, but this was not statistically significant.

Table 4.  Associations between mean importance score (1, high; 4, low) for all components and each of the four factors and patient characteristics
RespondentsAll componentsFactor*
1234
  1. Bold values significant at P < 0.05, Anova. *Factor 1, basics of prostate cancer care; 2, disease management; 3, psychosocial and physical well being; 4, self-help.

All respondents1.531.531.301.721.72
Age, years
≤701.431.421.261.541.69
71–771.501.481.301.701.68
≥781.651.721.341.911.76
Treatment type
Curative1.331.311.241.421.55
Palliative1.591.621.311.821.77
No treatment1.531.561.371.681.60
Depression
Depressed1.701.751.421.802.09
Not depressed1.401.481.281.701.64
Anxiety
Anxious1.571.481.361.731.73
Not anxious1.491.601.271.711.72
Social circumstances
With wife/partner1.501.501.301.671.66
Lives alone1.631.641.321.871.94
Duration of illness
<1 year1.521.571.321.611.69
≥1 year1.521.521.291.751.72

Multiple regression of the importance scores for each of the factors showed that although several variables were influential, the proportion of variance explained by them was very small. The factor with the greatest proportion of explained variance was ‘physical well being’ (F = 4.787, P < 0.01, adjusted R2 = 0.114) with explanatory factors comprising curative treatment, recent diagnosis and age (<78 years).

Cronbach's α of the 20 items gave satisfactory values for the whole 20-item scale and for factors 1–3, and borderline for Factor 4 (Table 5).

Table 5.  Internal consistency reliability analysis
ItemsReliability coefficient (Cronbach's a)
All 20 combined0.91
Factor 1: Basics of prostate cancer care0.87
Factor 2: Disease management0.80
Factor 3: Psychosocial & physical well being0.73
Factor 4: Self-help0.67

DISCUSSION

An extensive literature review showed the lack of a valid and reliable instrument for measuring the information needs of men with prostate cancer. In this study we sought to develop from first principles a cancer-specific tool capable of ascertaining such information needs, involving the target group from the earliest stages. A comprehensive list of information items was constructed and frequently reviewed to ensure that there were no omissions. There was a very high uptake, and patients found the tool to be understandable and easy to complete.

The study found that men with prostate cancer place considerable importance on a broad range of information needs, most of which are being inadequately met. Men varied in the importance that they placed on receiving different types of information, with associations detected between information needs and characteristics such as age; younger patients had greater information needs. However, multiple regression showed that very little variance could be explained by such variables. Although in part this reflected the relatively few patients sampled, it lends support to the need to tailor the information provided to the individual, their concerns and requirements.

The scale was tested for face, content and construct validity. Factor analysis showed that there was good construct validity, and no items were deleted from the 20-item scale as the factor loadings were all ≥ 0.506.

Determining the test-retest reliability of the scale was beyond the scope of this study, and furthermore the value of such testing is likely to be limited, given that information needs may be anticipated as changing over time. However, the reliability and internal consistency of the scale was high using Cronbach's α, especially for the whole 20-item scale and the ‘basics of prostate cancer care’ and ‘disease management’ factors. While these high α scores suggest potential redundancy, there were no significant differences in the value of internal consistency reliability if individual items were deleted. Indeed, deleting individual items did not lead to subsequent α values exceeding the overall α, which indicates that the 20-item scale could be left unchanged.

It was not possible to test for criterion validity, as there is no ‘reference standard’ research tool for measuring the information needs of patients with prostate cancer. However, there is much concordance between the items developed here and the areas of importance reported in prostate cancer studies in the USA and Canada [20,21].

LIMITATIONS

This was a pilot study with relatively few participants and thus caution is needed in generalizing the results. Although patients were recruited systematically, the general applicability of the findings to other settings needs further investigation. It cannot be assumed that patients with prostate cancer from different backgrounds would produce similar results.

The study sample excluded those diagnosed < 4 weeks previously and hence there may be further items that need to be included to reflect information needs in the period immediately after diagnosis. Furthermore, the analysis of the data showed a need to refine the wording of some items in the scale and to add to others. For example, it became clear that Item 1 (Information on prostate cancer in general) was too generic to be clinically useful. Nine patients left item 4 (What prostate cancer is) unanswered, and some comments were made to indicate that this item was not easily understood. During the study, additional themes were identified that suggested scope for improving the completeness and content validity of the scale; these included knowing about the disease prognosis, research in prostate cancer, accessing support services, and specific treatment side-effects.

IMPORTANCE OF CANCER-SPECIFIC INSTRUMENTS

Despite these limitations the scale offers a more valid means of assessing information needs in prostate cancer than has hitherto been reported. We recognize that patients are often unwilling or unable to articulate particular concerns, hence the need for strongly cancer-specific instruments to allow patients to report any such problems that their cancer or treatment might be causing [25]. For example, the modified version of the Toronto Informational Needs Questionnaire for patients with prostate cancer who are on hormonal therapy [24] has no items specifically addressing emergencies or warning signs, nor of the function of the prostate, the understanding of which has been identified as important and lacking amongst men with prostate cancer.

In parallel to this research we undertook a qualitative study to explore the information needs and care experiences of men with prostate cancer, the results of which will be published elsewhere. This confirmed that men wanted more information than they had received, whether it was positive or negative, and the perceived lack of information provided by health professionals left men feeling vulnerable, and likely to gather conflicting information from several sources.

In conclusion, many issues that are of concern to patients with prostate cancer appear not to be adequately recognized or addressed. Patients’ failure to articulate their information needs may lead clinicians to mistakenly believe that none exist. The instrument developed in the present study may provide a valid and reliable means of assessing information needs and concerns that relate to prostate cancer. This study provides a basis for further development of the scale, and testing of its impact on identifying needs and assessing the effectiveness of providing information in different formats (e.g. leaflets, consultations or videos); this will be the subject of future research.

ACKNOWLEDGEMENTS

We thank the clinicians and patients who gave their time to the project. Wolfram Jatsch was funded by the GP Unit, West Midlands Postgraduate Deanery. His work on this project formed his Masters thesis. Alison Pearce was funded by a T.P. Gunton Award from the BMA.

CONFLICT OF INTEREST

None declared. Source of funding: BMA.

Ancillary