Journal of Gerontological Nursing

The articles prior to January 2012 are part of the back file collection and are not available with a current paid subscription. To access the article, you may purchase it or purchase the complete back file collection here

A COMPARATIVE STUDY of Nurses' and Elderly Patients' Ratings of Pain and Pain Tolerance

Ingrid Bergh, RN, MSC; Björn Sjöström, PhD, RNT




Assessing and treating pain is one of the main tasks for geriatric nurses. Brattberg, Parker, and Thorslund (1996) found that among 537 individuals ages 77 to 98, 72.8% reported they had pain problems during the past year. Ferrell, Ferrell, and Osterweil (1990) found a similar high prevalence among people living in nursing homes, 70% of whom reported pain. There are only a few studies which focus on elderly people's pain (Melding, 1991). In a literature review of eight textbooks for geriatric nurses (encompassing 5,000 pages), Ferrell and Ferrell (1990) found that only 18 pages addressed the subject of pain alleviation. Closs (1994) considers pain in elderly people a neglected phenomenon. To improve the care of elderly patients who experience pain, nurses need more knowledge in this field.


Despite the fact that pain is a multidimensional problem, it is the intensity of the subjective experience that usually is measured when assessing pain and the effect of treatment (Herr & Mobily, 1993). One way to make the patients' pain more tangible to nurses is to use different types of measurement scales such as the Visual Analogue Scale (VAS) (Tiplady, Jackson, Maskrey, & Swift, 1998). Two studies conducted in postoperative care show that nurses' assessments often differ from patients' selfreports, and that patients' pain is underestimated by nurses (Sjöström, 1995; Zalon, 1993). Sjöström (1995) claims that a problem with pain assessment is that nurses' and patients' frames of reference differ, resulting in a deviation. Zalon (1993) found that nurses underestimated severe pain and overestimated mild pain, possibly because nurses expect that when patients report pain, the pain is severe. Zalon (1993) also reports that the underestimation of severe pain can be a way for nurses to handle situations in which they feel they have failed in successfully treating patients' pain. Sjöström (1995) showed that nurses systematically underestimate patients' pain and that this underestimation increased as patients rated their pain higher. Dalton (1989) posits that nurses with increased or continuous education spend more of their time considering the factors that influence the patients* expression of pain. Dudley and Holm (1984) found that nurses have a tendency to ascribe psychological distress to the patients rather than pain. Two other studies show that nurses attribute less pain to patients who lack a physical diagnosis (Half ens, Evers, & Abu-Saad, 1990; Taylor, Skelton, & Butcher, 1984).

Pain is defined from a point of view where the subjective experience is stressed. The International Association for the Study of Pain (IASP), Subcommittee on Taxonomy (1979), defines pain as:

An unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage (p. 250).

Consequently, pain is an experience and as such it cannot be shared with others in an objective way but must be communicated in a subjective way - the patient's way. McCaffery (1972) states:

Pain is whatever the experiencing person says it is, existing whenever he says it does (p. 8).

The fact that patients experience pain does not imply that patients are seeking relief from it. Bond (1981) notes that two types of pain thresholds exist, the lower of which is when pain occurs after an exposure to a sufficiently strong stimulus. The upper pain threshold is reached when it is impossible to increase the strength of the pain stimulus. Pain tolerance is the interval between these two thresholds. This interval varies widely between individuals because of personal and social factors (Bond, 1981). This way of describing the lower pain threshold is in line with the definition of IASP, Subcommittee on Taxonomy (1979), "The least stimulus intensity at which a subject perceives pain" (p. 252). However, IASP, Subcommittee on Taxonomy (1979) sets forth another meaning of the concept of pain tolerance, defined as:

The greatest stimulus intensity causing pain that a subject is prepared to tolerate (p. 252).

A variation of this definition is presented by Soafer (1992) in which she stresses the will to seek relief from pain:

Pain tolerance is the intensity of pain that an individual is willing to accept without seeking relief (p. 45). Soafer's (1992) definition of pain tolerance was used in this study. All definitions of pain thresholds and pain tolerance emphasize the individual experience, even if the concept is defined in different ways.

The purpose of this study was to compare elderly patients' and nurses' ratings of pain and pain tolerance.



The study was approved by the local Committee for Ethics in Medical Investigations, University of Gothenburg, Sweden, the senior physician at the hospital, and The Swedish Association of Health Officers. The study was conducted over a 3-month period. Nurses in nine geriatric units of a large university hospital in Sweden were informed and asked to participate. The participating nurses were required to work daytime shifts and at least part-time. In total, 66 nurses were asked and agreed to participate. However, only 59% (n = 39) of the nurses were eligible to participate in the study because there were only 39 patients available. Patients who were given information and asked to participate were admitted to the unit during the research period. A condition for participation was that they were able to understand verbal information. Inclusion criteria were arthrosis and nonpathological fracture. The initial selection consisted of 43 patients, 91% of whom participated in the study (n = 39, [arthrosis, ? = 7; nonpathological fracture, ? = 32]). Nine percent (n = 4) were unable to complete a VAS. These patients were not significantly older (mean age = 87, age range = 78 to 91) than those who completed the study (mean age = 81, age range = 67 to 96). Only one of the patients had previously used a VAS.


A VAS was used to assess the patients' pain, pain experienced earlier, and the patients' and nurses' concept of pain tolerance. The scale was a 10 cm straight line with end points marked "no pain" and "the worst pain imaginable." Patients were requested to mark a point which corresponded to the pain (Melzack & Katz, 1994). Several studies using the VAS have produced contradictory results (i.e., whether or not age affects the capacity to complete a pain estimation when using a VAS) (Choinière & Amsel, 1996; Herr & Mobily, 1993; Jensen, Karoly, & Braver, 1986; Kremmer, Hampton Atkinson & Ignelzi, 1981). Herr and Mobily (1991) said that detailed instructions are necessary for patients to understand the VAS and to be able to use it in a satisfactory way.

Figure 1. The relation between the nurses' ratings versus patients' rating of pain according to VAS (? = 39). Each data point on the scatterplot graph represents the patients' ratings of pain and the nurses' ratings of the patients' pain on the same occasion. Five observations are to be found in the origin of coordinates.

Figure 1. The relation between the nurses' ratings versus patients' rating of pain according to VAS (? = 39). Each data point on the scatterplot graph represents the patients' ratings of pain and the nurses' ratings of the patients' pain on the same occasion. Five observations are to be found in the origin of coordinates.

Patients were interviewed by the investigator and given the following instructions:

* Mark on the VAS the point which corresponds to your pain right now.

* Mark on the VAS the point which corresponds to how much pain you experienced before your hospitalization (i.e., under normal circumstances at home).

* Mark on the VAS the point which you consider corresponds to the least pain which should be treated in general.

Nurses followed the VAS instructions after they examined their patients regarding pain. The nurses were instructed to:

* Mark on the VAS the point which you consider corresponds to the patients' pain right now.

* Mark on the VAS the point which you consider corresponds to the least pain which should be treated in general.

The nurses also answered questions about experience, training, and length of service.


A structured interview was conducted by the investigator (LB.) with each of the selected patients because research has shown that elderly patients have an aversion to reporting pain (Ferrell et al., 1990; Yates, Dewar, & Fentiman, 1995). In general, the interviews were conducted in the patients' rooms, with the nurses absent, and lasted from 15 minutes to 1 hour. This procedure allowed the investigator an opportunity to give the patients accurate instructions about the VAS and to judge the fairness of the patients' ratings. In direct connection with the completed interview, the nurses assessed the patients' pain (using the VAS was not allowed). After completion of the pain assessment, the nurses completed a questionnaire.

Data Analysis

For participation, staff were chosen on the basis of being responsible for the care of the selected patients. Selection of patients was a convenience sample. Consistent with the statement, "Pain is whatever the experiencing person says it is, existing whenever he says it does" (McCaffery, 1972, p. 8), it was assumed that the patients' VAS ratings were valid. The estimations were transferred to a 10-cm VAS and measured to the nearest mm from "no pain" to the estimated point (Carlsson, 1983). Statistical calculations were made in the Statistical Package for the Social Sciences (SPSS) for Windows to reveal trends, keeping in mind the small sample. The mean value of the deviations was calculated to compare the estimations performed by nurses and patients for each group of nurses (e.g., with or without training) and for the group of patients (e.g., having experienced pain or not under normal circumstances at home). The differences between mean values were calculated using a two-tailed t test. To determine the relationship between the ratings of the patients and those of the nurses, a correlation analysis was used. If the nurses accurately assessed the patients' pain, the average deviation should have been nonsignificant and correlate highly with the patients' ratings. Despite the size of the sample, the authors chose to report the distribution as a percentage to facilitate comparisons.


Nurses' Ratings of the Patients' Pain

Nurses tended to rate pain lower (20 mm) than patients (30 mm) (p < .069) according to the VAS. The nurses' ratings also had only a moderate positive correlation with the patients' ratings (r = .39, ? < .02) (Figure 1).

The lower the amount of pain the patients rated, the less was the deviation from the assessments of the nurses. In the range 0 to 30 mm (n = 24), the deviation of the mean value in the case of the nurses was only +4 mm (t = -1.107, df= 2\p - not significant), but the correlation was low (r = .31, p = not significant). When patients rated their pain from 31 to 60 mm (n = 8), the mean value of the assessments of the nurses deviated - 10 mm (t = 1.010, df= 7, ? = not significant), and the correlation was moderate but not significant (r = .46, ? = not significant). Seven patients rated their pain from 61 to 100 mm on the VAS. The average deviation for the nurses was -52 mm (t = 4.362, df=6,p = .005), and there was virtually no correlation (r = -.10, p = not significant) between ratings.

Nurses' and Patients' Ratings of Pain Tolerance

Two nurses (n - 2) did not rate pain tolerance but commented in writing: "If the patient experiences pain and asks for treatment he or she should get it irrespective of the VAS score" and "This is individual."

Figure 2 shows that, on average, nurses (« = 37) rated pain tolerance at 14 mm, while patients (n = 39) rated it at 49 mm, resulting in a significant difference (t = 8.828, df= 36, ? < .000). Eighty-nine percent (n = 33) of nurses thought that pain rated at 30 mm according to VAS should be treated. Of the patients, only 18% (n = 7) held this view. Accordingly, there is a discrepancy between the ratings of nurses and patients regarding pain tolerance. When the assessed pain tolerance is subtracted from the rated pain, 21% (n = 8) of patients had passed their levels of pain tolerance, and one patient indicated the same level of experienced pain and pain tolerance. All patients who had marked their pain as equal to or greater than their pain tolerance also were considered by the nurses to be in pain. Seventy-seven percent (n = 30) rated their pain as being below their pain tolerance. Of this group, 12 patients were evaluated as being in pain by the nurses.

Figure 2. The distribution of the patients' (n = 39) and the nurses' (n = 37) ratings of pain tolerance according to VAS.

Figure 2. The distribution of the patients' (n = 39) and the nurses' (n = 37) ratings of pain tolerance according to VAS.

Nurses' Interpretations of Patients' Pain

The Table shows that of the patients (n = 29) who marked 1 to 100 mm on the VAS, 69% (? = 20) were regarded by the nurses as being in pain.

Impact of Training

Twenty-eight percent (w = 11) of the participating nurses stated they had had some kind of training in pain management beyond the standard nursing education. For 66% (n = 7) this training consisted of 1-day courses or occasional lectures, although one of the nurses reported 10 days of training in pain management. Nurses who had received some form of extra training in pain management (n = 11) had an average deviation of +3 mm (t = -0.423, df= 10,/» = .681) and had a correlation of r = .67 (p < .03) in their ratings. Those nurses with no extra training in pain management (n = 28) showed an average deviation of -14 mm (i = 2.317, df= 27, ? = .028), and they had a lower level of correlation with the patients' ratings (r = .34, ? < .08). Thus, nurses with extra training in pain management had a lower average deviation and a higher correlation with patient ratings, which suggests sensitivity to changes in the patients' pain.





Impact of Hours Worked Per Week

Full-time nurses (n = 28) had an average deviation from the patients in their ratings of -13 mm (t = 2.053, df= 27, ? = .050), and the correlation was moderate (r = .35, ? < .07). Nurses working part-time (n = 11) had an nonsignificant average deviation of -.2 mm (t = 0.026, df=ÍO,p = .980), and the correlation was stronger (r = .60, ? < .06). The nonsignificant average deviation and the higher correlation for the nurses working part-time suggests that they were better able to assess changes in the patients' pain than nurses who were employed full-time.

Impact of the Patients' Own Pain

Fifty-six percent (n = 22) of the patients reported that they had not encountered pain before they were hospitalized (i.e., under normal circumstances at home). For these patients, the nurses' ratings deviated -2 mm (t = .407, df = 21, ? = .688), and the correlation was r = .58 (p < .005). Forty-four percent (n = 17) reported that they had encountered pain before being hospitalized. The mean VAS score for this pain was 41 mm, and the cause of this pain varied (e.g., arthrosis, back pain). For this group, the mean value of the deviation was much higher (-19 mm) (r = 1.999, df=i6,p = .063, and the correlation was lower (r = .24, p = not significant).


The study suggests that nurses tend to overestimate patients' pain when patients mark "no pain" or "mild pain," and they underestimate the patients' pain when marked as "severe pain." This tendency to underestimate increases with the severity the patients mark on the VAS. Similar results have been found in other studies where nurses have rated postoperative patients' pain (Sjöström, 1995; Zalon, 1993). Sjöström (1995) suggests it is possible that nurses, through constant exposure to people in pain, acquire a frame of reference where the extreme value of "the worst pain imaginable" has changed, and this may be the reason for the lack of conformity in the assessments of nurses and patients. But why do nurses overestimate when patients mark "no pain" or "mild pain" and underestimate "severe pain"?

Sjöström (1995) claims that if patients rate their pain as 10 ("the worst pain imaginable") according to VAS, it is impossible for nurses to overestimate. The opposite applies when patients mark 0 ("no pain") when assessing pain (i.e., it is impossible for nurses to underestimate the patients' pain). When nurses and patients were asked to assess pain tolerance, significant differences were found (p < .000). One explanation of this large difference could be that many of the nurses have higher expectations when it comes to alleviating pain than the patients themselves.

Dalton (1989) found that approximately 50% of the nurses believe patients should be in a painless state; 30% of the nurses equated a painless state with "no pain." Also, a difference in frames of reference can exert an influence. For nurses as professionals, the extreme marking of "the worst pain imaginable" may not mean the same as it does to patients. Another important finding is that nurses may assume patients are in pain, although patients consider themselves within their personal pain tolerance.

It is possible that nurses can imagine the patients' pain, but differences may arise when patients and nurses assess pain from their respective frames of reference. Teske, Daut, and Cleeland (1983) show there is agreement between nurses and patients when stating the existence of pain, but there is less agreement when assessing the degree of pain. It is noteworthy that 21% (n = 8) of patients in this study felt they were in such pain that they needed treatment. In their investigations, Ferrei! et al. (1990) and Yates et al. (1995) found that elderly patients did not report pain. Seers (1987) reported in her study that 68% of the nurses believed if patients needed pain relief, they would ask for it, compared to 42% of the patients who expected the nurses to initiate pain relief.

The results of this study show that nurses with some training in pain management beyond basic nursing education tend to show less deviation and greater stability in their assessments, compared with nurses who have not had extra training in pain management. It is difficult to know whether training itself results in nurses assessing pain more correctly. It is somewhat unlikely that only 1 day of additional training would improve the capacity to assess pain. However, the higher correlations for the trained group suggest they are more sensitive to changes in patients' pain.

What is it that causes nurses to be more sensitive to patients' pain? Maybe a special interest in pain as a phenomenon is necessary to obtain knowledge beyond that provided by basic nursing education. Maybe it is an interest that improves these nurses' perceptions of patient pain. Even the length of service seems to influence the ability to assess pain. Nurses who worked part-time showed a lower deviation in their assessments and a higher correlation with patient ratings than nurses working fulltime. Zalon (1993) feels that the reason nurses underestimate severe pain could be because they had not noticed patients' pain at an early stage. A failure of this kind could cause frustration and feelings of powerlessness (Soafer, 1992). It could be that the nurses working part-time are more emotionally relaxed. This may improve their empathy and perceptions. The nurses' training in pain management and length of service and patients' pain prior to hospitalization tend to influence the nurses' ability to assess patients' pain. The nurses' average assessment deviation was lower, while the correlation was higher for patients who had no pain prior to hospitalization, compared with those who had previously experienced pain.

What is it that influences nurses in their assessments of these two groups of patients, and why is the result different? Patients who experience pain for a long period of time develop "chronic pain behavior" (Forrest, 1995, p. 17). Jacox (1979) explains that many people who suffer from chronic pain develop control strategies to cope with it. Jacox (1979) states that one problem when assessing patients who suffer from chronic pain is that they are expected to show the same behaviors as patients with acute pain. Taylor et al. (1984) similarly found that nurses ascribe significantly less pain to those who suffer from chronic pain than to those who suffer from acute pain.


The high prevalence of pain among elderly patients underscores the need for well-defined guidelines and routines in geriatric care to detect, assess, treat, and document patients' pain. These routines should be developed in such a way that they can be adapted easily to patients' individual needs. Concepts such as pain intensity, pain tolerance, and the purpose of treatment should be based on patients' frames of reference and not on those of nurses. As Scott and Huskisson (1975) state:

Measurement of pain must always be subjective since pain is a subjective phenomenon, only the patient can therefore measure its severity (p. 184).

To alter the guidelines and routines associated with pain management, two areas need to be improved. First, nurses must acquire increased knowledge regarding pain and pain management. Second, nurses must change their practices regarding pain management. It is important for nurses to update their knowledge. This study suggests a deficiency in training and education in pain management. One way of achieving both increased knowledge and changes in nursing practice concerning pain management could be the use of research-based knowledge to improve clinical practice. Dufault, Bielecki, Collins, and Willey (1995) found in their study that nurses who participated in research projects involving practical problems change their nursing practice. They changed their way of assessing, treating, and documenting their patients' pain.


The purpose of this study was to compare elderly patients' and nurses' ratings of pain and pain tolerance. Data were collected through structured interviews with the patients. The attending nurses completed a questionnaire after conducting a pain assessment. Independent of each other, patients and nurses were asked to rate on a VAS when pain should be treated (pain tolerance) and pain intensity. The VAS has been used both by patients and nurses. The initial selection consisted of 43 patients; however, 9% n = 4) were unable to complete the VAS. These patients were not significantly older than those who completed the study (n = 39). The results shows that nurses tend to overestimate mild pain and underestimate severe pain. Nurses rated pain tolerance significantly lower than patients. The results also suggest that nurses with training beyond basic nursing education tend to assess patients' pain more accurately than those without additional training. For patients who reported that they had pain prior to hospitalization, the nurses' pain ratings showed a higher agreement than for those who reported that they did not have pain before being hospitalized. At the time of the interviews, 21% (n = 8) of patients felt that their pain was so great they needed treatment. Those patients also were recognized by the attending nurses as being in pain. To improve elderly patients' pain management, practicing nurses must collaborate with researchers to develop specific empirical research nursing knowledge within geriatric pain management. This research-based knowledge should be incorporated into nurses' clinical practice regarding pain management. Specific guidelines must be developed for the assessment, treatment, and documentation of elderly patients' pain.


  • Bond, M.R. (1981). Pain, its nature, analysis and treatment. Singapore: Longman Group Limited.
  • Brattberg, G., Parker, M.G., & Thorslund, M. (1996). The prevalence of pain among the oldest old in Sweden. Pain, 67, 29-34.
  • Garlsson, A.M. (1983). Assessment of chronic pain. I: Aspects of the reliability and validity of die visual analogue scale. Pain, 16, 87-101.
  • Choinière, M., & Amsel, R. (1996). A visual analogue thermometer for measuring pain intensity. Journal of Pain and Symptom Management, 11(5), 299-311.
  • Closs, SJ. (1994). Pain in elderly patients: A neglected phenomenon? Journal of Advanced Nursing, 19, 1072-1081.
  • Dalton, J. A. (1989). Nurses' perceptions of their pain assessment skills, pain management practices, and attitudes toward pain. Oncology Nursing Forum, 16(2), 223-231.
  • Dudley, S.R., & Holm, K. (1984). Assessment of the pain experience in relation to selected nurse characteristics. Pain, 18, 179-186.
  • Dufault, M.A., Bielecki, C., Collins, E., & Willey, C. (1995). Changing nurses' pain assessment practice: A collaborative research utilization approach. Journal of Advanced Nursing, 21, 634-645.
  • Ferrell, B.R., & Ferrell, B.A. (1990). Easing the pain. Geriatric Nursing, 11, 175-178.
  • Ferrell, B.A., Ferrell, B.R., & Osterweil, D. (1990). Pain in the nursing home. Journal of the American Geriatria Society, 38, 409-414.
  • Forrest, J. (1995). Assessment of acute and chronic pain in older adults. Journal of Gerontological Nursing, 27(10), 15-20.
  • Halfens, R., Evers, G., & Abu-Saad, H. (1990). Determinants of pain assessment by nurses. International Journal of Nursing Studies, 27(1), 43-49.
  • Herr, K. A., & Mobily, RR. (1991). Complexities of pain assessment in the elderly. Clinical considerations. Journal of Gerontological Nursing, 17(4), 12-19.
  • Herr, K.A., & Mobily, P.R. (1993). Comparison of selected pain assessment tools for use with the elderly. Applied Nursing Research, 6(1), 39-46.
  • International Association for the Study of Pain, Subcommittee on Taxonomy. (1979). Pain terms: A list with definitions and notes on usage. Pain, 6, 249-252.
  • Jacox, A.K. (1979). Assessing pain. American Journal of Nursing, 79, 895-904.
  • Jensen, M.P., Karoly, P., & Braver, S. (1986). The measurement of clinical pain intensity: A comparison of six methods. Pain, 27, 117-126.
  • Kremmer, E., Hampton Atkinson, J., Sc Ignelzi, RJ. (1981). Measurement of pain: Patient preference does not confound pain measurement. Pain, 10, 241-248.
  • McCaffery, M. (1972). Nursing management of the patient with pain. Philadelphia: Lippincott.
  • Melding, P.S. (1991). Is there such a thing as geriatric pain? Pain, 46, 119-121.
  • Melzack, R., & Katz, J. (1994). Pain measurement in persons in pain. In P.D. Wall & R. Melzack (Eds.), Textbook of pain (3rd ed., pp. 337-351). Edinburgh: Churchill Livingstone.
  • Scott, J., & Huskisson, E.C. (1975). Graphic representation of pain. Pain, 2, 175-184.
  • Seers, CJ. (1987). Perceptions of pain. Nursing Times, 83(4%), 37-39.
  • Sjöström, B. (1995). Assessing acute postoperative pain. Assessment strategies and quality in relation to clinical experience and professional role. Göteborg, Sweden: Acta Universitatis Gothoburgensis.
  • Soafer, B. (1992). Pain: A handbook fornurses (2nd ed.). London, England: Chapman & Hall.
  • Taylor, A.G., Skelton, J.A., & Butcher, J. (1984). Duration of pain condition and physical pathology as determinants of nurses' assessments of patients in pain. Nursing Research, 33(1), 4-8.
  • Teske, K., Daut, R.L., & Cleeland, CS. (1983). Relationships between nurses' observations and patients' self-reports of pain. Pain, 16, 289-296.
  • Tiplady, B., Jackson, S.H.D., Maskrey, V.M., Swift, CG. (1998). Validity and sensitivity of visual analogue scales in young and older healthy subjects. Age and Ageing, 27, 63-66.
  • Yates, P., Dewar, A., & Fentiman, B. (1995). Pain: The views of elderly people living in long-term residential care settings. Journal of Advanced Nursing, 21, 667-674.
  • Zalon, M.L. (1993). Nurses' assessment of postoperative patients' pain. Pain, 54, 329334.




Sign up to receive

Journal E-contents