- Email: [email protected]
Measuring musculoskeletal pain by questionnaires: The manikin versus written questions
European Journal of Pain 14 (2010) 335–338
Contents lists available at ScienceDirect
European Journal of Pain journal homepage: www.EuropeanJournalPain.com
Measuring musculoskeletal pain by questionnaires: The manikin versus written questions Linda H.J. van den Hoven a,c, Kees J. Gorter b, H. Susan J. Picavet a,* a
Centre for Prevention and Health Services Research (PZO, pb 101), National Institute of Public Health and the Environment, P.O. Box 1, 3720 BA Bilthoven, The Netherlands Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, The Netherlands c University Medical Centre Utrecht, The Netherlands b
a r t i c l e
i n f o
Article history: Received 7 January 2009 Received in revised form 22 May 2009 Accepted 16 June 2009 Available online 20 August 2009 Keywords: Musculoskeletal pain Population-based surveys Questionnaire Manikin Pain drawing
a b s t r a c t A picture of a human figure (manikin) on which pain can be indicated can be used to measure musculoskeletal pain in self administered or web-based questionnaires. In this paper we present an analysis of the agreement between pain reported on a manikin and pain reported using written questions as assessed in the follow-up questionnaire of the Dutch population-based Musculoskeletal Conditions and Consequences Cohort (DMC3-study). Both a manikin and extensive questions on pain were included and the agreement between the two measures was studied for nine pain locations. For a similar pain definition – pain lasting at least a week – the manikin gave slightly higher prevalences than the written questions. Around three quarter of those who reported pain on the written questionnaire also indicated pain on the manikin on the same anatomical location. There were no differences in the percentage of agreement by sex, age group or level of education, except for a lower percentage of agreement of the manikin among the elderly for pain in the lower extremities, neck and shoulder and among the lower level of education group for neck pain. Almost 6% of the participants reported pain according to the questions only and more than 10% (especially men) reported pain on the manikin only. We concluded that a manikin gives similar findings on prevalence of pain as written questions and could therefore be a good alternative for written questions only. Ó 2009 European Federation of International Association for the Study of Pain Chapters. Published by Elsevier Ltd. All rights reserved.
1. Introduction
2. Materials and methods
Manikins are widely used to measure pain prevalence on different areas of the human body (Croft et al., 1996; Ohnmeiss, 2000; Jinks et al., 2001; Lacey et al., 2003; Carnes et al., 2007). A manikin is a simple line drawing of an outline of the human body, which provides a method that does not limit pain reporting to any particular location (Schierhout and Myers, 1996). A manikin can be seen as an attractive component that gives some variation in questionnaire surveys and for which cultural and linguistic differences might be of less influence. Pain manikins are widely used to report on pain, but it is not well known how pain reported on manikins corresponds with pain reported on traditional (written) questions. The aim of the present study is to assess the agreement between the manikin and a self administered questionnaire on musculoskeletal pain for nine anatomical locations and specified for age, sex and level of education.
Follow-up data of the Dutch population-based Musculoskeletal Complaints and Consequences Cohort study (DMC3-study) (details in: Picavet and Schouten, 2003) were analyzed: the follow-up questionnaire was send to 2752 adults who had given permission in the baseline questionnaire (n = 3664) 6 months before, 2389 (87%) responded, slightly more women than men. A 24-pages full-colour questionnaire was used; it consisted of general questions and health questions. Pages of five different colours corresponded with one of the following five anatomical areas: (1) neck, shoulder or upper back, (2) elbow or wrist/hand, (3) lower part of the back, (4) hip or knee and (5) ankle/foot. The anatomical areas included three, two or one anatomical location(s), in total nine anatomical locations. Every coloured area started with a screening question on pain during the past 6 months and those screening positive were asked to answer all questions of the relevant colour. The questionnaire ended with the manikin with both anterior and posterior aspects of the body. Participants were asked to shade where they have had pain from muscles, joints, bones or (part of the) back for at least 1 week during the last month. It was
* Corresponding author. Tel.: +31 30 2743063; fax: +31 30 2744407. E-mail address: [email protected] (H.S.J. Picavet).
1090-3801/$36.00 Ó 2009 European Federation of International Association for the Study of Pain Chapters. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ejpain.2009.06.002
336
L.H.J. van den Hoven et al. / European Journal of Pain 14 (2010) 335–338
emphasized that these complaints should not be related to influenza or menstruation. Sociodemographic characteristics were gathered in the baseline questionnaire. In order to compare the written questions with the manikin, the prevalence of current pain lasting longer than 1 week was calculated for nine anatomical locations. The manikins were scored using a transparent plastic template divided into different body parts: the posterior aspect was divided into 23 body parts, the anterior aspect into 21, see Fig. 1. We combined these data on different body parts to represent the anatomical locations of the written questions as close as possible. This was done as follows: neck: F(ront) 18 and B(ack) 18; shoulder: F1, F5, B1, B5; upper back: B22, B23, B24; elbow: F2, F6, B2, B6; wrist/hand: F3, F4, F7, F8, B3, B4, B7, B8; lower Back: B25, B26; Hip: F9, F13, B9, B13; knee: F10, F14, B10, B14; ankle/foot: F12, F16, B12, B16. The association between the manikin and questions was expressed in a percentage of agreement, which was calculated as
the sensitivity of the manikin with the written questions as the gold standard, with 95% confidence interval. Two disconcordant groups – persons who did have complaints according to the questions, but did not according to the manikin and vice versa – were identified. All analyses of data were performed using SPSS version 12.0.1. 3. Results Of all responders 53% had pain for at least 1 week according to the written questions. Pain prevalence measured by the manikin was slightly higher at 60%. The prevalence patterns by sex, age or level of education were very similar for pain measured by questions or manikin (Table 1). The mean agreement between manikin with the written questions was 75%, ranging from 62% for the upper back to 81% for ankle/foot (Table 2). For men and women the agreement differed by
Fig. 1. Template manikin.
337
L.H.J. van den Hoven et al. / European Journal of Pain 14 (2010) 335–338 Table 1 Prevalence (%) of pain lasting more than 1 week according to questions and to manikin for nine anatomical locations by sociodemographic characteristics.
Neck, question Neck, manikin Shoulder, question Shoulder, manikin Higher back, question Higher back, manikin Elbow, question Elbow, manikin Wrist-hand, question Wrist-hand, manikin Lower back, question Lower back, manikin Hip, question Hip, manikin Knee, question Knee, manikin Ankle-foot, question Ankle-foot, manikin
Total
By sex
By age
By education
(n = 2389)
Men (n = 1065)
Women (n = 1324)
Age 25–45 (n = 808)
Age 46–65 (n = 926)
Age > 65 (n = 655)
Low (n = 351)
Average (n = 1455)
High (n = 560)
17.8 16.4 17.9 21.4 7.0 12.2 6.8 9.7 12.7 14.1 22.0 25.6 9.6 15.6 15.8 19.7 10.0 12.1
14.9 13.7 14.7 18.3 5.7 9.6 6.4 8.5 9.0 10.6 21.1 25.2 6.7 12.8 13.9 18.8 8.6 11.3
20.2 19.1 20.4 24.2 8.2 14.9 7.2 11.0 15.7 17.4 22.7 26.7 12.0 18.1 17.3 21.2 11.1 13.4
14.3 16.6 16.5 19.7 6.3 14.6 4.5 6.6 10.3 10.9 19.1 24.5 3.8 12.0 9.3 14.7 6.4 9.0
18.0 21.4 20.7 25.8 8.7 13.1 9.1 13.6 14.9 17.0 24.8 28.8 11.1 16.0 17.1 21.2 11.6 14.0
17.1 13.8 15.7 17.4 5.6 7.9 6.6 8.0 12.6 13.8 21.6 22.4 14.8 19.4 21.9 23.8 12.1 13.3
18.6 11.6 16.6 20.9 7.8 8.4 9.5 10.7 14.8 17.7 25.7 24.3 12.9 18.7 19.9 25.8 10.7 13.3
19.3 18.4 20.1 22.8 7.7 13.1 7.2 10.7 12.6 13.2 21.4 25.6 9.9 15.3 6.0 20.1 9.2 11.5
13.6 14.3 13.6 18.6 5.2 12.5 4.6 6.8 12.3 14.5 17.7 26.3 7.2 14.3 12.7 15.1 12.0 13.4
Table 2 Percentage of agreement; percentage of pain as measured with the question that was indicated on the manikin. Complaints of:
Men (%)
95% CI
Women 95% CI (%)
Age 25– 45 (%)
95% CI
Age 46– 65 (%)
95% CI
Age > 65 (%)
95% CI
Education 95% CI Low
Education Average
95% CI
Education High
95% CI
Neck Shoulder Higher back Elbow Wrist-hand Lower back Hip Knee Ankle-foot
70.3 75.6 61.7 75.0 72.9 78.8 76.1 78.9 83.5
63.1–77.4 68.9–82.4 49.4–74.0 64.7–85.3 64.0–81.8 73.5–84.2 66.1–86.0 72.3–85.5 75.9–91.1
68.2 80.6 62.6 75.8 77.7 78.5 72.2 78.3 80.1
80.0 79.7 78.4 75.0 79.5 83.1 83.9 84.0 88.5
72.7–87.3 72.9–86.5 67.1–89.7 60.9–89.1 70.8–88.2 77.2–89.0 70.9–96.8 75.7–92.3 79.8–97.1
67.5 83.2 55.0 76.2 75.9 78.4 75.5 81.5 82.2
61.0–74.1 77.8–88.5 44.1–65.9 67.1–85.3 68.8–83.1 73.1–83.8 67.1–83.8 75.5–87.6 75.0–89.5
60.0 69.3 55.6 74.4 73.2 74.1 67.7 72.3 75.6
50.8–69.2 60.3–78.3 39.3–71.8 61.4–87.5 63.6–82.8 66.8–81.4 58.4–77.1 65.0–79.7 66.1–85.2
48.4 75.4 37.0 69.7 84.3 71.6 71.1 77.9 75.7
72.0 77.9 64.0 76.9 73.6 78.6 69.9 80.2 81.2
62.1–82.0 67.8–88.1 49.1–78.8 60.1–93.8 61.2–86.1 68.8–88.5 56.2–83.6 68.7–91.7 65.9–96.5
75.0 85.5 79.3 76.9 76.8 84.6 87.5 73.2 85.1
55.5–94.5 64.7–106.3 46.9–111.7 43.2–110.6 56.1–97.5 67.9–101.3 58.5–116.5 53.3–93.1 63.0–107.2
62.6–73.8 75.9–85.3 53.4–71.8 67.2–84.4 72.0–83.4 73.9–83.2 65.2–79.1 72.9–83.7 73.7–86.6
location, but on average it had the same agreement of 75%. The percentage of agreement of the manikin was reduced notably in the oldest age group (>65 years) and the low educated group. In the elderly percentage of agreement was less on lower extremities, neck and shoulder. In the lower educated group the agreement was reduced: 48% for neck pain. In the higher educated group the agreement was higher, ranging from 73% to 88%. The percentage of agreement did not improve when the definition of pain according to the questions was limited to more severe pain in terms of pain intensity or longer duration of pain (data not shown). Furthermore, pain lasting less then 1 week was often also indicated on the manikin: for instance 55 responders reported shoulder pain lasting less than 1 week on the written questions and 36 of them indicated this pain on the manikin. Almost 6% (5.7%) reported pain on the written questions but not on the manikin, and almost twice as much reported pain on the manikin (10.4%) but not on the written questions. The ‘manikin only’ group represents relatively more men. 4. Discussion Comparing pain measured by a manikin or by written questions showed that the manikin gave slightly higher prevalences for pain and that the agreement was fairly high: three quarter of all pain mentioned on the written questions was also mentioned on the manikin, with only a few differences on location of pain and sociodemographic subgroups in the population.
36.2–60.7 64.3–86.6 18.8–55.3 54.0–85.4 74.3–94.3 62.2–81.0 57.9–84.4 68.1–87.8 61.9–89.5
The relative higher pain prevalence found by the manikin – and also found when studied for men and women, by age group or by eductional level – may indicate that the threshold for reporting pain is lower for the manikin than for the questions. Earlier studies, however, provide not a univocal picture: for knee pain higher prevalences were found using a question compared to a blank manikin. (Jinks et al., 2001). For shoulder pain a simple question gave lower prevalences compared to a blank manikin (Pope et al., 1997), but a question with a pre-shaded manikin gave the highest prevalence. Using a pre-shaded manikin for neck and upperlimb pain resulted also in relative higher prevalences (Lacey et al., 2003). The mean percentage of agreement found was 75%, indicating that 75% of the pain as measured with the questions was indicated on the manikin. Overall there was no difference in ‘agreement’ of the manikin by location of pain, accept for the upper back. The phrase ‘upper back’, i.e. the part of the back between neck and lower back, is probably a less specific anatomical location compared to the other locations. An explanation for lower agreement in the low educated and oldest age group might be the lower cognitive capacities in both groups. Lacey et al. (2003) compared a blank manikin to be shaded with a question illustrated by a pre-shaded manikin to measure upper limb and neck pain. They found an agreement of 88%. Those who disagreed consisted of two groups (2% blank manikin only, 10% pre-shaded) with no differences for men and women and with highest proportion of disagreements in the oldest age group. The study of Jinks et al. was focused on knee pain. Of all participants
338
L.H.J. van den Hoven et al. / European Journal of Pain 14 (2010) 335–338
who shaded the manikin and answered the knee pain question (n = 78), 95% had also answered yes to the knee pain question. (Jinks et al., 2001). The large non-corresponding group with only shading of the manikin and the higher prevalence figures found with the manikin in our study can suggest that participants have a higher propensity to use a manikin to indicate their pain. This is not inconceivable because the visual aspect can make the manikin more appealing to respond to. The following limitations should be taken into account. First, the pain definition described in the instruction for the manikin could not be replicated exactly with the written questions: the definition of current pain lasting at least a week was the best we could get. The resulting bias may have resulted in higher prevalences in the manikin. Second, the additional instruction for the manikin was to not shade pain related to influenza or menstruation. This could have reduced the pain prevalence, yet the prevalence figures were higher for most of the locations on the manikin, except for the neck. Third, with the relative high non-response in the baseline study and some additional non-response in the follow-up we could have introduced some selection bias. However, although the estimation of the prevalence of pain for the total reference population may be biased (Picavet and Schouten, 2003) it is not expected that this should affect an analysis between two measures within the study. A fourth limitation is the preset order of the questionnaire; after reading and answering many questions about musculoskeletal pain the manikin had to be shaded. It is not known if percentage of agreement figures will change when the questionnaire starts with the manikin. Finally, the locations described in the questions had to be translated to locations on the manikin. For some locations this was a bit problematic. For example the hip consisted in the manikin also of the upper leg. Taking all these limitations into
account we think the analyses show that the manikin gives globally the same results as the written questions. For epidemiological surveys the manikin can represent an attractive change in a questionnaire which consists mainly of words and sentences. Acknowledgements The Dutch population-based Musculoskeletal Complaints & Consequences Cohort (DMC3-study) was financially supported by the Ministry of Health, Welfare and Sport of the Netherlands and the National Institute of Public Health and the Environment and carried out in collaboration with Statistics Netherlands. References Carnes D, Parsons S, Ashby D, Breen A, Foster NE, Pincus T, et al. Chronic musculoskeletal pain rarely presents in a single body site: results from a UK population study. Rheumatology 2007;46:1168–70 [Oxford]. Croft P, Burt J, Schollum J, Thomas E, Macfarlane G, Silman A. More pain, more tender points: is fibromyalgia just one end of a continuous spectrum? Ann Rheum Dis 1996;55:482–5. Jinks C, Lewis M, Ong BN, Croft P. A brief screening tool for knee pain in primary care. 1. Validity and reliability. Rheumatology 2001;40:528–36 [Oxford]. Lacey RJ, Lewis M, Sim J. Presentation of pain drawings in questionnaire surveys: influence on prevalence of neck and upper limb pain in the community. Pain 2003;105:293–301. Ohnmeiss DD. Repeatability of pain drawings in a low back pain population. Spine 2000;25:980–8. Picavet HS, Schouten JS. Musculoskeletal pain in the Netherlands: prevalences, consequences and risk groups, the DMC(3)-study. Pain 2003;102:167–78. Pope DP, Croft PR, Pritchard CM, Silman AJ. Prevalence of shoulder pain in the community: the influence of case definition. Ann Rheum Dis 1997;56:308–12. Schierhout GH, Myers JE. Is self-reported pain an appropriate outcome measure in ergonomic–epidemiologic studies of work-related musculoskeletal disorders? Am J Ind Med 1996;30:93–8.
Contents lists available at ScienceDirect
European Journal of Pain journal homepage: www.EuropeanJournalPain.com
Measuring musculoskeletal pain by questionnaires: The manikin versus written questions Linda H.J. van den Hoven a,c, Kees J. Gorter b, H. Susan J. Picavet a,* a
Centre for Prevention and Health Services Research (PZO, pb 101), National Institute of Public Health and the Environment, P.O. Box 1, 3720 BA Bilthoven, The Netherlands Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, The Netherlands c University Medical Centre Utrecht, The Netherlands b
a r t i c l e
i n f o
Article history: Received 7 January 2009 Received in revised form 22 May 2009 Accepted 16 June 2009 Available online 20 August 2009 Keywords: Musculoskeletal pain Population-based surveys Questionnaire Manikin Pain drawing
a b s t r a c t A picture of a human figure (manikin) on which pain can be indicated can be used to measure musculoskeletal pain in self administered or web-based questionnaires. In this paper we present an analysis of the agreement between pain reported on a manikin and pain reported using written questions as assessed in the follow-up questionnaire of the Dutch population-based Musculoskeletal Conditions and Consequences Cohort (DMC3-study). Both a manikin and extensive questions on pain were included and the agreement between the two measures was studied for nine pain locations. For a similar pain definition – pain lasting at least a week – the manikin gave slightly higher prevalences than the written questions. Around three quarter of those who reported pain on the written questionnaire also indicated pain on the manikin on the same anatomical location. There were no differences in the percentage of agreement by sex, age group or level of education, except for a lower percentage of agreement of the manikin among the elderly for pain in the lower extremities, neck and shoulder and among the lower level of education group for neck pain. Almost 6% of the participants reported pain according to the questions only and more than 10% (especially men) reported pain on the manikin only. We concluded that a manikin gives similar findings on prevalence of pain as written questions and could therefore be a good alternative for written questions only. Ó 2009 European Federation of International Association for the Study of Pain Chapters. Published by Elsevier Ltd. All rights reserved.
1. Introduction
2. Materials and methods
Manikins are widely used to measure pain prevalence on different areas of the human body (Croft et al., 1996; Ohnmeiss, 2000; Jinks et al., 2001; Lacey et al., 2003; Carnes et al., 2007). A manikin is a simple line drawing of an outline of the human body, which provides a method that does not limit pain reporting to any particular location (Schierhout and Myers, 1996). A manikin can be seen as an attractive component that gives some variation in questionnaire surveys and for which cultural and linguistic differences might be of less influence. Pain manikins are widely used to report on pain, but it is not well known how pain reported on manikins corresponds with pain reported on traditional (written) questions. The aim of the present study is to assess the agreement between the manikin and a self administered questionnaire on musculoskeletal pain for nine anatomical locations and specified for age, sex and level of education.
Follow-up data of the Dutch population-based Musculoskeletal Complaints and Consequences Cohort study (DMC3-study) (details in: Picavet and Schouten, 2003) were analyzed: the follow-up questionnaire was send to 2752 adults who had given permission in the baseline questionnaire (n = 3664) 6 months before, 2389 (87%) responded, slightly more women than men. A 24-pages full-colour questionnaire was used; it consisted of general questions and health questions. Pages of five different colours corresponded with one of the following five anatomical areas: (1) neck, shoulder or upper back, (2) elbow or wrist/hand, (3) lower part of the back, (4) hip or knee and (5) ankle/foot. The anatomical areas included three, two or one anatomical location(s), in total nine anatomical locations. Every coloured area started with a screening question on pain during the past 6 months and those screening positive were asked to answer all questions of the relevant colour. The questionnaire ended with the manikin with both anterior and posterior aspects of the body. Participants were asked to shade where they have had pain from muscles, joints, bones or (part of the) back for at least 1 week during the last month. It was
* Corresponding author. Tel.: +31 30 2743063; fax: +31 30 2744407. E-mail address: [email protected] (H.S.J. Picavet).
1090-3801/$36.00 Ó 2009 European Federation of International Association for the Study of Pain Chapters. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ejpain.2009.06.002
336
L.H.J. van den Hoven et al. / European Journal of Pain 14 (2010) 335–338
emphasized that these complaints should not be related to influenza or menstruation. Sociodemographic characteristics were gathered in the baseline questionnaire. In order to compare the written questions with the manikin, the prevalence of current pain lasting longer than 1 week was calculated for nine anatomical locations. The manikins were scored using a transparent plastic template divided into different body parts: the posterior aspect was divided into 23 body parts, the anterior aspect into 21, see Fig. 1. We combined these data on different body parts to represent the anatomical locations of the written questions as close as possible. This was done as follows: neck: F(ront) 18 and B(ack) 18; shoulder: F1, F5, B1, B5; upper back: B22, B23, B24; elbow: F2, F6, B2, B6; wrist/hand: F3, F4, F7, F8, B3, B4, B7, B8; lower Back: B25, B26; Hip: F9, F13, B9, B13; knee: F10, F14, B10, B14; ankle/foot: F12, F16, B12, B16. The association between the manikin and questions was expressed in a percentage of agreement, which was calculated as
the sensitivity of the manikin with the written questions as the gold standard, with 95% confidence interval. Two disconcordant groups – persons who did have complaints according to the questions, but did not according to the manikin and vice versa – were identified. All analyses of data were performed using SPSS version 12.0.1. 3. Results Of all responders 53% had pain for at least 1 week according to the written questions. Pain prevalence measured by the manikin was slightly higher at 60%. The prevalence patterns by sex, age or level of education were very similar for pain measured by questions or manikin (Table 1). The mean agreement between manikin with the written questions was 75%, ranging from 62% for the upper back to 81% for ankle/foot (Table 2). For men and women the agreement differed by
Fig. 1. Template manikin.
337
L.H.J. van den Hoven et al. / European Journal of Pain 14 (2010) 335–338 Table 1 Prevalence (%) of pain lasting more than 1 week according to questions and to manikin for nine anatomical locations by sociodemographic characteristics.
Neck, question Neck, manikin Shoulder, question Shoulder, manikin Higher back, question Higher back, manikin Elbow, question Elbow, manikin Wrist-hand, question Wrist-hand, manikin Lower back, question Lower back, manikin Hip, question Hip, manikin Knee, question Knee, manikin Ankle-foot, question Ankle-foot, manikin
Total
By sex
By age
By education
(n = 2389)
Men (n = 1065)
Women (n = 1324)
Age 25–45 (n = 808)
Age 46–65 (n = 926)
Age > 65 (n = 655)
Low (n = 351)
Average (n = 1455)
High (n = 560)
17.8 16.4 17.9 21.4 7.0 12.2 6.8 9.7 12.7 14.1 22.0 25.6 9.6 15.6 15.8 19.7 10.0 12.1
14.9 13.7 14.7 18.3 5.7 9.6 6.4 8.5 9.0 10.6 21.1 25.2 6.7 12.8 13.9 18.8 8.6 11.3
20.2 19.1 20.4 24.2 8.2 14.9 7.2 11.0 15.7 17.4 22.7 26.7 12.0 18.1 17.3 21.2 11.1 13.4
14.3 16.6 16.5 19.7 6.3 14.6 4.5 6.6 10.3 10.9 19.1 24.5 3.8 12.0 9.3 14.7 6.4 9.0
18.0 21.4 20.7 25.8 8.7 13.1 9.1 13.6 14.9 17.0 24.8 28.8 11.1 16.0 17.1 21.2 11.6 14.0
17.1 13.8 15.7 17.4 5.6 7.9 6.6 8.0 12.6 13.8 21.6 22.4 14.8 19.4 21.9 23.8 12.1 13.3
18.6 11.6 16.6 20.9 7.8 8.4 9.5 10.7 14.8 17.7 25.7 24.3 12.9 18.7 19.9 25.8 10.7 13.3
19.3 18.4 20.1 22.8 7.7 13.1 7.2 10.7 12.6 13.2 21.4 25.6 9.9 15.3 6.0 20.1 9.2 11.5
13.6 14.3 13.6 18.6 5.2 12.5 4.6 6.8 12.3 14.5 17.7 26.3 7.2 14.3 12.7 15.1 12.0 13.4
Table 2 Percentage of agreement; percentage of pain as measured with the question that was indicated on the manikin. Complaints of:
Men (%)
95% CI
Women 95% CI (%)
Age 25– 45 (%)
95% CI
Age 46– 65 (%)
95% CI
Age > 65 (%)
95% CI
Education 95% CI Low
Education Average
95% CI
Education High
95% CI
Neck Shoulder Higher back Elbow Wrist-hand Lower back Hip Knee Ankle-foot
70.3 75.6 61.7 75.0 72.9 78.8 76.1 78.9 83.5
63.1–77.4 68.9–82.4 49.4–74.0 64.7–85.3 64.0–81.8 73.5–84.2 66.1–86.0 72.3–85.5 75.9–91.1
68.2 80.6 62.6 75.8 77.7 78.5 72.2 78.3 80.1
80.0 79.7 78.4 75.0 79.5 83.1 83.9 84.0 88.5
72.7–87.3 72.9–86.5 67.1–89.7 60.9–89.1 70.8–88.2 77.2–89.0 70.9–96.8 75.7–92.3 79.8–97.1
67.5 83.2 55.0 76.2 75.9 78.4 75.5 81.5 82.2
61.0–74.1 77.8–88.5 44.1–65.9 67.1–85.3 68.8–83.1 73.1–83.8 67.1–83.8 75.5–87.6 75.0–89.5
60.0 69.3 55.6 74.4 73.2 74.1 67.7 72.3 75.6
50.8–69.2 60.3–78.3 39.3–71.8 61.4–87.5 63.6–82.8 66.8–81.4 58.4–77.1 65.0–79.7 66.1–85.2
48.4 75.4 37.0 69.7 84.3 71.6 71.1 77.9 75.7
72.0 77.9 64.0 76.9 73.6 78.6 69.9 80.2 81.2
62.1–82.0 67.8–88.1 49.1–78.8 60.1–93.8 61.2–86.1 68.8–88.5 56.2–83.6 68.7–91.7 65.9–96.5
75.0 85.5 79.3 76.9 76.8 84.6 87.5 73.2 85.1
55.5–94.5 64.7–106.3 46.9–111.7 43.2–110.6 56.1–97.5 67.9–101.3 58.5–116.5 53.3–93.1 63.0–107.2
62.6–73.8 75.9–85.3 53.4–71.8 67.2–84.4 72.0–83.4 73.9–83.2 65.2–79.1 72.9–83.7 73.7–86.6
location, but on average it had the same agreement of 75%. The percentage of agreement of the manikin was reduced notably in the oldest age group (>65 years) and the low educated group. In the elderly percentage of agreement was less on lower extremities, neck and shoulder. In the lower educated group the agreement was reduced: 48% for neck pain. In the higher educated group the agreement was higher, ranging from 73% to 88%. The percentage of agreement did not improve when the definition of pain according to the questions was limited to more severe pain in terms of pain intensity or longer duration of pain (data not shown). Furthermore, pain lasting less then 1 week was often also indicated on the manikin: for instance 55 responders reported shoulder pain lasting less than 1 week on the written questions and 36 of them indicated this pain on the manikin. Almost 6% (5.7%) reported pain on the written questions but not on the manikin, and almost twice as much reported pain on the manikin (10.4%) but not on the written questions. The ‘manikin only’ group represents relatively more men. 4. Discussion Comparing pain measured by a manikin or by written questions showed that the manikin gave slightly higher prevalences for pain and that the agreement was fairly high: three quarter of all pain mentioned on the written questions was also mentioned on the manikin, with only a few differences on location of pain and sociodemographic subgroups in the population.
36.2–60.7 64.3–86.6 18.8–55.3 54.0–85.4 74.3–94.3 62.2–81.0 57.9–84.4 68.1–87.8 61.9–89.5
The relative higher pain prevalence found by the manikin – and also found when studied for men and women, by age group or by eductional level – may indicate that the threshold for reporting pain is lower for the manikin than for the questions. Earlier studies, however, provide not a univocal picture: for knee pain higher prevalences were found using a question compared to a blank manikin. (Jinks et al., 2001). For shoulder pain a simple question gave lower prevalences compared to a blank manikin (Pope et al., 1997), but a question with a pre-shaded manikin gave the highest prevalence. Using a pre-shaded manikin for neck and upperlimb pain resulted also in relative higher prevalences (Lacey et al., 2003). The mean percentage of agreement found was 75%, indicating that 75% of the pain as measured with the questions was indicated on the manikin. Overall there was no difference in ‘agreement’ of the manikin by location of pain, accept for the upper back. The phrase ‘upper back’, i.e. the part of the back between neck and lower back, is probably a less specific anatomical location compared to the other locations. An explanation for lower agreement in the low educated and oldest age group might be the lower cognitive capacities in both groups. Lacey et al. (2003) compared a blank manikin to be shaded with a question illustrated by a pre-shaded manikin to measure upper limb and neck pain. They found an agreement of 88%. Those who disagreed consisted of two groups (2% blank manikin only, 10% pre-shaded) with no differences for men and women and with highest proportion of disagreements in the oldest age group. The study of Jinks et al. was focused on knee pain. Of all participants
338
L.H.J. van den Hoven et al. / European Journal of Pain 14 (2010) 335–338
who shaded the manikin and answered the knee pain question (n = 78), 95% had also answered yes to the knee pain question. (Jinks et al., 2001). The large non-corresponding group with only shading of the manikin and the higher prevalence figures found with the manikin in our study can suggest that participants have a higher propensity to use a manikin to indicate their pain. This is not inconceivable because the visual aspect can make the manikin more appealing to respond to. The following limitations should be taken into account. First, the pain definition described in the instruction for the manikin could not be replicated exactly with the written questions: the definition of current pain lasting at least a week was the best we could get. The resulting bias may have resulted in higher prevalences in the manikin. Second, the additional instruction for the manikin was to not shade pain related to influenza or menstruation. This could have reduced the pain prevalence, yet the prevalence figures were higher for most of the locations on the manikin, except for the neck. Third, with the relative high non-response in the baseline study and some additional non-response in the follow-up we could have introduced some selection bias. However, although the estimation of the prevalence of pain for the total reference population may be biased (Picavet and Schouten, 2003) it is not expected that this should affect an analysis between two measures within the study. A fourth limitation is the preset order of the questionnaire; after reading and answering many questions about musculoskeletal pain the manikin had to be shaded. It is not known if percentage of agreement figures will change when the questionnaire starts with the manikin. Finally, the locations described in the questions had to be translated to locations on the manikin. For some locations this was a bit problematic. For example the hip consisted in the manikin also of the upper leg. Taking all these limitations into
account we think the analyses show that the manikin gives globally the same results as the written questions. For epidemiological surveys the manikin can represent an attractive change in a questionnaire which consists mainly of words and sentences. Acknowledgements The Dutch population-based Musculoskeletal Complaints & Consequences Cohort (DMC3-study) was financially supported by the Ministry of Health, Welfare and Sport of the Netherlands and the National Institute of Public Health and the Environment and carried out in collaboration with Statistics Netherlands. References Carnes D, Parsons S, Ashby D, Breen A, Foster NE, Pincus T, et al. Chronic musculoskeletal pain rarely presents in a single body site: results from a UK population study. Rheumatology 2007;46:1168–70 [Oxford]. Croft P, Burt J, Schollum J, Thomas E, Macfarlane G, Silman A. More pain, more tender points: is fibromyalgia just one end of a continuous spectrum? Ann Rheum Dis 1996;55:482–5. Jinks C, Lewis M, Ong BN, Croft P. A brief screening tool for knee pain in primary care. 1. Validity and reliability. Rheumatology 2001;40:528–36 [Oxford]. Lacey RJ, Lewis M, Sim J. Presentation of pain drawings in questionnaire surveys: influence on prevalence of neck and upper limb pain in the community. Pain 2003;105:293–301. Ohnmeiss DD. Repeatability of pain drawings in a low back pain population. Spine 2000;25:980–8. Picavet HS, Schouten JS. Musculoskeletal pain in the Netherlands: prevalences, consequences and risk groups, the DMC(3)-study. Pain 2003;102:167–78. Pope DP, Croft PR, Pritchard CM, Silman AJ. Prevalence of shoulder pain in the community: the influence of case definition. Ann Rheum Dis 1997;56:308–12. Schierhout GH, Myers JE. Is self-reported pain an appropriate outcome measure in ergonomic–epidemiologic studies of work-related musculoskeletal disorders? Am J Ind Med 1996;30:93–8.