An observational study of pain self-management strategies and outcomes: does type of pain, age, or gender, matter?

2.1 Eligibility criteria

To be eligible, participants had to be pharmacists between 18 and 80 years of age who had an email address listed with the State Board of Pharmacy. To be included in the study, participants had to indicate that they had acute or chronic pain within the past 5 years. Pharmacists were excluded if they did not have a computer or on-line access or if the email address was not valid. This study was approved by the University Human Subjects Protection Program.

2.2 Sample

This study used a convenience sample of pharmacists recruited via email. Because the prevalence of pain in this population group was not known, the final sample size was estimated to be between 200 and 600 participants.

2.3 Instrument

To develop the questionnaire, studies using population-based questionnaires of pain were reviewed, topics and questions were adapted from these questionnaires [25], [26], [27], [28], [29], [30], [31]. Acute pain was defined as pain experienced only occasionally, most of the time an individual was pain free. Chronic pain was defined as pain experienced almost every day or as recurring pain for three of the past 6 months [32].

The questionnaire was constructed in sections that included demographics, pain characteristics, strategies used to manage pain, and outcomes. Demographic questions included age, gender, marital status, ethnicity, primary practice site, number of years of active practice, current employment status, disability status, and health status. Questions on pain characteristics included intensity of pain before treatment, level at which pain is tolerable, pain management goal, and sleep pain level (level of pain at which they are able to sleep), measured on an 11-point (0–10) numeric rating scale (NRS). Participants were also asked to identify the cause of their pain from a list of choices (e.g. accident/injury, arthritis, etc.).

To identify treatment strategies, participants were asked if they used prescription NSAIDS or opioids, or if they used OTC medications including aspirin, acetaminophen, OTC NSAIDS, or herbal remedies as well as non-medical strategies including hot/cold packs, massage, rest, or avoiding activities. Participants were asked to identify their primary method of managing pain with choices of medical strategies (defined as medications and other treatments that require access to the health care system), non-medical strategies (defined as any strategy that was not a medication or did not require a prescription), or a combination.

2.4 Outcomes

Outcome measures included level of pain after treatment, using an 11-point NRS rating scale (0–10), percent pain relief from use of all strategies, satisfaction rated as not at all to very satisfied, and confidence in their ability to manage their pain therapy (0–10).

The questionnaire was administered using an on-line questionnaire program (Qualtrics©) [33]. The link to the questionnaire was included in the recruitment email. Responses were automatically entered into a database in an anonymous manner. The questionnaire was pilot tested by the research group.

2.5 Data analysis

Data were analyzed using SPSS© (version 24) [34]. The analysis strategy was to compare the acute and chronic pain groups and if substantial differences were found, the acute and chronic pain groups would be analyzed separately. Effect sizes (Cohen’s d) for comparisons were calculated to assist in the determination of clinical importance of identified differences, less than 0.2 was considered small, 0.2–0.79, moderate, and greater than or equal to 0.8, large [35]. Continuous data were analyzed using a t-test, and categorical data, using a χ²-test. Hierarchical logistic regression was used to identify independent predictors of gender and age. Demographic characteristics were entered in step 1, pain characteristics in step 2, pain management strategies in step 3, and outcomes in step 4. The a priori alpha level was 0.05.

3.1 Acute pain

The demographic characteristics for the acute pain group stratified by gender and age are shown in the left portion of Table 3. Women were more likely to rate their health status as excellent and to be married (p<0.040). Women and men did not differ in disability (p=0.465). As would be expected, when stratified by age, years of practice, age, and employment status, were different (p<0.05).

Pain characteristics stratified by gender and age for the acute pain group are shown in the left portion of Table 4. There were no differences between men and women on pain intensity before treatment, level of pain tolerable, goal level of pain, and sleep pain level (ES’s<0.23). Also, there were no differences in confidence (ES<0.10). However, women were more likely to report migraine as the cause of their pain (ES=0.33). Related to age, the only difference was that older participants were more likely to report arthritis as a cause of their pain (ES=0.69).

As shown in the left portion of Table 5, there were no differences in pain management strategies for men and women with acute pain and only two differences related to age. Older participants were more likely to use herbal remedies (ES=0.38) and younger participants were more likely to use rest as a strategy to manage acute pain (ES=0.46).

3.2 Chronic pain

The demographic characteristics for the chronic pain group stratified by gender and age are shown in the right portion of Table 3. Women were younger, had practiced for less time, and were less likely to be married than men (p<0.004), however, men and women did not differ in disability or health status. When stratified by age, older participants were more likely to be male and to report a disability (ES=0.89).

Pain characteristics for the chronic pain group are shown in the right portion of Table 4. With respect to gender, men and women did not differ on baseline pain intensity, tolerable level of pain, goal level, and level of pain at which they were able to sleep (ES’s<0.23). Women were more likely to report migraine (ES=0.60) whereas men, to report arthritis (ES=1.00). Men and women did not differ in confidence.

With respect to age, older participants with chronic pain had somewhat greater pain intensity before treatment (ES=0.35), a higher level of tolerable pain (ES=0.47), and a higher goal level of pain (ES=0.40). Older participants with chronic pain also were more likely to report arthritis and migraine (ES=0.40 & 0.50).

There were several differences in the use of pain management strategies in the chronic pain group (right portion of Table 5). With respect to gender, men were more likely to use prescription NSAIDS (ES=0.37), and women were more likely to use hot/cold packs (ES=0.34) and massage (ES=0.42). Although the use of herbal remedies was low, women were more likely to use herbal remedies than men (ES=0.34). With respect to age differences, younger participants were more likely to use massage (ES=0.53) while older participants were somewhat more likely to report use of prescription NSAIDs (0.32).

The majority of participants (62%) in both the acute and chronic pain groups identified a combination strategy including both medical and non-medical strategies as their preferred overall strategy (See Table 5). Within the chronic and acute pain groups, men and women and younger and older participants did not differ on their preference for a combination strategy (ES’s=0.01–0.09).

3.3 Outcomes

The outcomes of pain management are shown at the bottom of Table 5. In the acute pain group (shown on the left side of Table 5), there were no significant differences on level of pain post treatment, percent relief from use of all strategies, and satisfaction, based on gender or age. There was also no difference in outcomes based on gender in the chronic pain group. However, for the chronic pain group, the effect sizes for post-treatment level of pain (ES=0.29) and percent relief (ES=0.34) were borderline moderate for older versus younger participants. Also notable was the substantial variability across all strata in post-treatment levels of pain (coefficient of variation ranged from 68% to 100%) and percent relief from all strategies (coefficient of variation ranged from 25% to 32%).

3.4 Independent predictors of gender and age

The hierarchical logistic regression identified few independent predictors of gender and age related to use of management strategies or outcomes. In the acute pain group, use of herbal products (p=0.030) and rest (p=0.004) were predictive of female gender. Use of NSAIDS was marginally predictive of male gender in both the acute and chronic pain groups (p=0.056 & 0.131, respectively). No management strategies were predictive of age in the chronic pain group, however, tolerable level of pain was predictive (p=0.003).

4.1 Scientific inference and a non-representative sample

To infer whether or not there is an association between gender or age and use of pain management strategies and outcomes, there must be reasonable certainty that gender and age are not confounded with some other variable (e.g. level of education). In this study, confounders were controlled by including participants whose level of education, occupation, access to health care services, and socio-economic status were similar. That is, the sample was restricted which is a recognized method of controlling confounders [17], [18]. Hence, the findings that there were substantial differences between the acute and chronic pain groups, and that associations between gender or age and pain management strategies and outcomes were small, should represent valid findings that were not confounded with socioeconomic variables. However, because the study sample is not representative, the findings do not have statistical generalizability and do not represent the associations that might be found in a random sample of a population where confounding from socioeconomic variables would be expected. Additional studies examining pain in socioeconomically homogenous populations would help clarify the age and gender relationships between pain and demographic characteristics, pain characteristics, pain management strategies, and outcomes.

4.2 Clinical implications

Based on the assumption that pain management could be improved if both patients and providers understand differences related to type of pain, acute or chronic, and gender and age, the findings from this study suggest that the differences based on type of pain are substantial while those based on gender are minimal although there might be some meaningful differences based on age. Hence, from a clinical perspective, the differences between acute and chronic pain may be important while those related to gender and age, not as important. In addition, as noted in the IOM report [1], and as found in this study, variability in pain management strategies and outcomes are large. Hence, the most useful clinical strategy may be to focus on individualization of management strategies irrespective of age and gender.

4.3 Implications for research in pain management

The IOM noted in their report that acute and chronic pain are often not differentiated [1]. However, because of the number and size of the differences found between the acute and chronic pain groups in this socioeconomically homogenous sample, we recommend that individuals and their reported pain experience and management be separated by pain group, acute and chronic. Our findings also suggest that the results of laboratory models of pain may have limited generalizability to chronic pain since laboratory models involving healthy human subjects (e.g. a study using capsaicin [38] or hypertonic saline injections [39] to produce pain) represent acute pain. Our findings also support the conclusion of Racine et al. [24] that a new paradigm would be helpful for studying gender differences in acute and chronic pain. Meaningful differences may exist within population subgroups in which the subculture or cultural norms related to gender vary from that of the general population [4] however, the findings would be specific to that subpopulation. Meaningful differences based on age seem possible in chronic pain. In univariate analysis, there were six differences (ES>0.30 & <0.51) in pain characteristics and four differences (ES>0.28 <0.54) in pain management or outcomes between younger and older participants. Given that average age of the older group in this study was low (mean=62 years), there could be additional differences if more participants were older.

4.4 Issues with the recall of pain intensity

Recall of pain intensity has been shown to be an issue in using self-reported data to study pain management. In general, individuals tend to rate their pain higher when recalling the pain than when it is rated contemporaneously. Recalled baseline pain ratings have been shown to be higher following orthopedic surgery [40], following emergency department visits for minor trauma [41], for non-surgery individuals in a hospital rehabilitation program [42], and in individuals with arthritis or fibromyalgia [43]. However, no differences were found in the recall of migraine pain [44]. Several explanations have been proposed to explain observations of bias. One proposed explanation is response bias. Response bias refers to the use of different scales when recalling pain intensity versus contemporaneous ratings of pain [40]. Another explanation may be that recall of pain intensity is a different construct than is the contemporaneous measure of pain [43]. Because this study used a questionnaire and recalled pain intensity ratings, recall bias should be kept in mind when interpreting the results. That the effect of recall bias may differ based on the type of pain also may have increased variability.