Epidemiology of pain in back and extremities in rural population: A community-based estimation of age- and sex-specific prevalence, distribution, duration and intensity of pain, number of painful sites and seasonality of pain during twelve months in rural Gadchiroli, India

Background Population-based estimates of the burden of pain in back and extremities (PBE) by sex, age, intensity, seasonality and site are lacking from rural India. Methods Two villages were randomly selected from a cluster of 39 villages in Gadchiroli district in India. All residents’≥20 years of age were surveyed in January 2010 by trained surveyors by making household visits. Information on PBE in the 12 months prior to survey was obtained using a structured, pretested questionnaire. Results The 12-month period prevalence of PBE was 75% (95% confidence interval CI = 72.54-77.73) in men and 91% (95% CI = 88.66-92.13) in women. The prevalence of PBE in the participants >50 years was 94% while that in the age group 20 to 50 years was 79% (P < 0.05). The site with the highest prevalence of pain was low back (women 80%, men 59%). The mean number of painful sites per person was 5.42 (95% CI = 5.17-5.67) in women, 3.68 (95% CI = 3.45-3.90) in men, 3.89 (95% CI = 3.71-4.07) in participants aged 20 to 50 years and 6.48 (95% CI = 6.11-6.85) in those >50 years. Among participants across the age and sex groups, the prevalence of mild pain was higher than severe pain at all the anatomical sites. Among various seasons, the highest prevalence of pain was in the rainy season (14%). Conclusion The prevalence and the number of painful sites were higher among women and in those >50 years of age. The public health interventions for PBE need to focus on these two high risk groups.

However, population-based data on the epidemiology of the burden of back and musculoskeletal pain from rural India, focusing on communities which almost exclusively are agrarian and rely on manual labour is lacking, thereby also limiting the possibility of developing appropriate interventions.
This study, estimating the age and sex specific burden of pain in back and extremities (PBE) in the adult (≥20 years) population in rural Gadchiroli in the Maharashtra state in India fills this gap. The burden was measured as, age-and sex-specific a) 12-month period prevalence, b) site specific prevalence, c) number of painful sites per adult, d) duration of pain, e) the intensity of pain and f) seasonality of pain in the rural adult population in Gadchiroli, India, over a period of 12 months.

Study design and sample size
This study was nested in a population-based, cross-sectional survey of the prevalence of PBE in rural Gadchiroli. The study setting, study design, detail method of sample size calculation, method of village selection including the eligibility criteria are described in detail elsewhere [10].
Questionnaire development, training, quality control and data collection A questionnaire in vernacular language (Marathi) was developed to interview the participants and record the following information about: a) episodes of pain in back and extremities at different body sites, b) duration of each pain episode and c) intensity of pain at each site in the 12 months preceding January 2010. The questionnaire was pilot tested in villages that were not part of the study and in the rural clinic of SEARCH. Community health workers (CHWs) were trained in administering the questionnaire and the data were collected from 1 January 2010 to 25 January 2010, details of which are described elsewhere [10].

Statistical methods
A database was constructed for data entry using FOX PRO Version 2.0 (Microsoft Inc, Redmond, Washington, USA). The data were double entered, validated and checked for inconsistencies. Descriptive statistics included mean, medians and ranges for continuous variables and proportions for categorical variables were calculated. The age and sex specific 12-month period prevalence and the number of painful sites were estimated with their associated 95% confidence intervals (CI). Student's t test was used for comparison of means. Differences between proportions were assessed using Chi-sqaure test. We followed STROBE guidelines for the reporting of observational studies. Analyses were conducted using Stata 10.0 (Stata Corp, College Station, Texas, USA).

Ethical approval
The research followed the tenets of the Declaration of Helsinki. Ethical approval for this nested study was granted as part of the main study, by the Institutional Ethical Committee of SEARCH formed according to the guidelines by the Indian Council for Medical Research. Consent was obtained first at the cluster level in the study villages 15 days before starting the survey. The community leaders (Village Council Leaders and members, school-teacher and presidents of microfinance self-help groups) were explained the purpose and scope of the study including the benefits to the villagers (availability of referral care in SEARCH clinic and the care through a village clinic). Informed written consent in vernacular language in a standard format was obtained from individual participants after explaining the nature and benefits of the study. The benefits provided during the study included free consultation by spine surgeons and rheumatologists in a clinic conducted in the same village at a later date. For those who needed further evaluation, laboratory investigations, as well as imaging with Magnetic Resonance Imaging (MRI) and x-ray including transport were provided free of cost. For patients needing pharmacotherapy and physiotherapy, these services were provided free of cost and for those needing surgical interventions, such services were provided at significantly subsidized costs. The CHWs discussed these benefits using a printed pamphlet.

RESULTS
The total population of the two villages was 3735 out of which 2535 (67.9%) were adults ≥20 years of age and were eligible to participate in the study (Figure 1). Of these, 2259 (89%) were interviewed, 276 (11%) were either absent from the village (migrated for work) or unable to communicate due to very old age or disability. Total 1101 men (49%) and 1158 (51%) women participated in the study. Proportion of illiteracy was higher in women (55%) participants than men (22.2%). The demographic characteristics of the participants are described in Table 1.  Figure 2, and Table  S2 in the Online Supplementary Document).

Number of painful sites
The mean number of painful sites per woman participant 5.42 (95% CI = 5.17-5.67) was significantly higher than in men 3.68 (95% CI = 3.45-3.90). The mean number of painful sites per participant according to age also significantly increases from 2.81 (95% CI = 2.58-3.05) in the age group 20 to 30 years to 6.62 (95% CI = 6.10-7.14) in the age group of >60 years (Figure 3, Figure 4, and Tables S3 and S4 in the Online Supplementary Document). The prevalence of 1 to 5 painful sites was comparable in men and women (50.3% and 49.8% respectively) but the prevalence of 6 to 10 painful sites was significantly higher in women (27.5%) than men (17.2%). The prevalence of zero painful sites was more than twice in men (24.8%) than women (9.5%), while the prevalence of more than 10 painful sites was almost twice in women (13.2%) compared to men (7.7%) ( Table S5 in the Online Supplementary Document). Overall, both the prevalence as well as number of painful sites were higher in women and older age group suggestive of a sex and age-related gradient.  Figure 5).

Duration of pain
Overall, the prevalence of acute (1 to 42 days) and chronic (more than 84 days) pain was significantly higher than the prevalence of sub-acute (43 to 84 days) pain in both men and women at all the anatomical sites. The prevalence of acute pain was higher in men than in women at all the anatomical sites (Table 5). In men, the prevalence of acute pain at neck and superior extremity (23% and 24% at respective sites) was higher than that of chronic pain (8% and 18% at respective sites) while the prevalence of acute (15%) and chronic (16%) thoracic pain was almost equal. In women, the prevalence of chronic pain was higher at all the anatomical sites. In the older participants, the prevalence of chronic pain was higher than acute pain at all the anatomical sites, whereas in the      *Categories are multiple and overlapping. †Any one of shoulder, arm, elbow, forearm, wrist, hand + fingers, trapezius/scapula. ‡Any one of hip/buttocks, groin, thigh, knee, leg/calf, ankle, heel, foot + digits. younger participants, specifically at neck (22%) and the superior extremity (22%), the prevalence of acute pain was more than chronic pain (14% and 18% at respective sites).

Intensity of pain according to sex and age and seasonality of pain
In both men and women, as well as participants across the age groups, the prevalence of mild pain was higher than severe pain at all the anatomical sites (Figure 6, Table S5 in the Online Supplementary Document).
Similarly, for all the anatomical sites, the highest number of participants reported pain in all three seasons. Among those who reported pain during only a specific season, the prevalence of pain was higher in rainy season (Figure 7, Table S6 in the Online Supplementary Document).

DISCUSSION
The five anatomical sites with the highest prevalence of pain, in both men and women as well as across the age groups were low back, knee, neck, leg and thoracic region. This may be due to higher load bearing activities of rural communities involving these sites. The high prevalence of neck pain can be due to the practice of carrying heavy load (firewood, earth) on head for long distance, especially by women.
The prevalence of acute pain was higher in men and younger participants than women and older for all the anatomical sites. The association between sudden unexpected maximum efforts such as bending, twisting, lifting with low back pain is known [11] and may be contributing to higher prevalence of acute back pain in men who are generally engaged in heavy manual labour. The higher prevalence of acute pain at neck (23%) and superior extremity (24%) in men, was probably due to heavy load bearing and digging. Therefore, specific interventions are needed to address the acute pain in this group. Apart from these exceptions, chronicity of pain was the norm corroborating with previously published literature [12]. Constant wear and tear of the bone and muscular frame because of physical hard labour, poor ergonomic postures, and certain nutritional factors may be the possible risk factors, but further studies are needed to identify precise causative pathways.
The prevalence of mild pain was higher than severe pain at all the anatomical sites. This is a hopeful sign, possibly suggestive of good coping mechanisms of the communities. Therefore, simple pain relief measures may yield results in reducing prevalence of PBE. The higher prevalence of low back pain, especially in the rainy season may be due to the bent position women adopt for long hours during planting saplings of paddy and harvesting as well as the heavy load carried by and manual work undertaken by men. Hence ergonomic interventions involving appropriate, low cost technology to reduce the physical burden of paddy cultivation or similar seasonal manual labour specific to the region may be explored as an intervention.
We compared the results from our study with those from other COPCORD studies. The higher prevalence of pain in low back than knee in both the sexes and in different age groups was similar to results of COPCORD studies from China [13], Indonesia [14], aborigines in Australia [15] and Bangladesh (8)    the findings of the Indian COPCORD series in which the prevalence of knee pain was higher than low back [7,9]. In general, women have a higher burden of acute and nonfatal chronic morbidities [16], especially temporary and persistent pain [17]. Our study also reported higher prevalence of pain in women at all the anatomical sites. This corroborated with other COPCORD studies [18][19][20][21], even though it differed from smaller set of studies reporting higher prevalence of back pain in men [22][23][24] or equal between sexes [25]. The highest prevalence of pain during the rainy season in our study differs from the Indian COPCORD study [19] in which the highest prevalence was in winter. This could be due to higher proportion of agrarian population in our study which is engaged in manual agricultural work the most during monsoon.
The factors affecting higher prevalence of PBE in our population would include higher prevalence of osteoporosis in women with increasing age compared to men [1]. This may have may have contributed to the higher prevalence of musculoskeletal pains in the women, especially in the older age group. Prevalence of pain at particular sites, especially at knee may be higher in women due to higher proportion of osteoarthritis, for which female sex, low body weight and lifting are known risk factors [26]. Nevertheless, this does not fully explain the difference in prevalence of knee pain between women and men, as the other risk factors such as heavy lifting and alcohol abuse [26] are also common in men. It has been reported that the causative factors of low back pain in African population differed from those in the Western population [27]. Similarly, therefore, the Indian population may have certain specific risk factors especially considering the less muscle mass and height as well as role of genetic factors [27].
To the best of our knowledge, this is the first study from rural India reporting in detail the population-based epidemiology of PBE including prevalence according to age and sex, the number of painful sites, duration, intensity as well as the seasonality of pain. The previous studies on epidemiology of PBE were primarily from peri-urban and relatively affluent rural areas [28], restricted to a certain type of labour population only, such as drill worker, cashew worker or jute worker [29][30][31], or were not population based [32]. However, it is possible that the epidemiology of PBE would vary according to different rural parts of India due to regional differences, the nature of occupational work communities are engaged in, socioeconomic status and access to care. In this aspect, the participants in this study can be considered more representative of the larger agrarian Indian communities with its significant dependence on manual labour.
This study, nested in the larger study to identify the overall 12 months period prevalence had several strengths which included random selection of the two study villages, high participation rate of the participants (89%) and data collection by CHWs with more than 15 years of experience using a well-tested, structured and robust questionnaire with rigorous quality control. These lend confidence to the estimates obtained. The key limitations of our study are the possibility of recall loss in reporting of pains from preceding 12 months and routine treatment of pains using aspirin by CHWs in these villages. Nevertheless, both these limitations would have resulted in underestimation and not overestimation of the burden of the problem.

CONCLUSION
This population-based study from rural Gadchiroli describes the epidemiology of pain in back and extremities, identifies a significant gender and age related burden and underlines the chronic nature of the problem. It also provides clues to the different target populations (younger vs older and men vs women) for public health intervention. Further larger population based studies from different parts of rural India are needed to identify the regional estimates of the epidemiological variation in the profile of PBE as well as risk factors to understand the causation and explore intervention for PBE in rural Indian communities.