Research article | Open | Open Peer Review | Published:
Primary care consultation, hospital admission, sick leave and disability pension owing to neck and low back pain: a 12-year prospective cohort study in a rural population
BMC Musculoskeletal Disordersvolume 7, Article number: 66 (2006)
Neck and low back pain are common musculoskeletal complaints generating large societal costs in Western populations. In this study we evaluate the magnitude of long-term health outcomes for neck and low back pain, taking possible confounders into account.
A cohort of 2,351 Swedish male farmers and rural non-farmers (40–60 years old) was established in 1989. In the first survey, conducted in 1990–91, 1,782 men participated. A 12-year follow-up survey was made in 2002–03 and 1,405 men participated at both times. After exclusion of 58 individuals reporting a specific back diagnosis in 1990–91, the study cohort encompassed 1,347 men. The health outcomes primary care consultation, hospital admission, sick leave and disability pension were assessed in structured interviews in 2002–03 (survey 2). Symptoms and potential confounders were assessed at survey 1, with the exception of rating of depression and anxiety, which was assessed at survey 2. Multiple logistic regression generating odds ratios (OR) with 95% confidence intervals (95% CI) was performed to adjust the associations between reported symptoms and health outcomes for potential confounders (age, farming, workload, education, demand and control at work, body mass index, smoking, snuff use, alcohol consumption, psychiatric symptoms and specific back diagnoses during follow up).
Of the 836 men reporting current neck and/or low back pain at survey 1, 21% had had at least one primary care consultation for neck or low back problems, 7% had been on sick leave and 4% had disability pension owing to the condition during the 12 year follow up. Current neck and/or low back pain at survey 1 predicted primary care consultations (OR = 4.10, 95% CI 2.24–7.49) and sick leave (OR = 3.22, 95% CI 1.13–9.22) after potential confounders were considered. Lower education and more psychiatric symptoms were independently related to sick leave. Lower education and snuff use independently predicted disability pension.
Few individuals with neck or low back pain were on sick leave or were granted a disability pension owing to neck or low back problems during 12 years of follow up. Symptoms at baseline independently predicted health outcomes. Educational level and symptoms of depression/anxiety were important modifiers.
Pain and discomfort from the neck and the low back are very common complaints in the general population. The majority of middle aged people have experienced such symptoms for shorter or longer periods [1–4]. In Sweden, musculoskeletal disorders are still the most common reason for sick leave and disability pension, although psychiatric illnesses are on the increase . In spite of the fact that neck and low back pain are so frequent, have been extensively researched and carry large societal costs, we still have limited knowledge about the causes as well as the course of these conditions.
Associations between both physical workload and psychosocial factors and neck pain [6, 7] and low back pain [8–10] have been reported in numerous studies. However, there is still not sufficient evidence to support causal relations regarding these factors . Psychological factors seem to be important to the development of these conditions over time, to how pain is conceived and handled in terms of care seeking and sick listing . There is limited data on the course of low back pain  and almost no data on the course of neck pain in populations over longer periods. Although the majority of pain episodes are mild, many have recurrent episodes and a minority develop persistent chronic pain . How health care consumption, sick leave and disability pensions develop over a longer time span in relation to previously reported symptoms is insufficiently studied. In view of the major individual suffering and societal costs for sick leave compensation and disability pensions in Sweden and comparable countries, studies of prospective data on these matters are a pressing matter .
In the research project "Health through work" we have followed a cohort of initially occupationally active middle-aged farmers and rural non-farmers over 12 years. At the baseline survey in 1990–91 the farmers generally reported more musculoskeletal symptoms than non-farmers . Despite this, farmers had not sought medical care more often than non-farmers and had reported less sick listing. The two groups differ considerably in physical and psychosocial background characteristics and therefore the cohort is suitable for studies on musculoskeletal health outcomes. The purpose of the present study was to describe the frequency of primary health care consultations, hospital admissions, sick leave and disability pensions due to neck and low back pain during 12 years of follow up in relation to reported symptoms at baseline taking potential confounders into account.
In 1989, a study cohort of 2,351 male farmers and non-farmers (aged 40–60 years), from nine different municipalities across Sweden, was created for prospective studies of primarily salutogenes. The municipalities were selected to cover known morbidity gradients across the country and to include areas with various types of farming. The farmers were identified using the Swedish Register of Farming, and occupational activity in farming (>25 hours per week) was thoroughly checked. The non-farmers were sampled from the national population register and matched to the farmers by sex, age and residential area. The non-farmers had to be occupationally active in other than farming. The sampling procedure has been described in more detail previously [16–19].
The study cohort was invited to participate in an extensive health survey in 1990–91, and 1,782 men (75.8%) participated in this baseline survey (survey 1). In 2002–03, 1,589 (67.6%) men participated in a second similar survey (survey 2). A description of the study cohort and reasons for non-participation is given in table 1. Both surveys were performed with specially trained personnel traveling to the various areas, and they were carried out as separate research projects and not part of any ordinary health program. The participation rate was somewhat higher among the farmers than among the non-farmers although the setting was the same, and the groups were not addressed differently. Fifty-eight men reporting a specific back diagnosis at survey 1 (rheumatoid arthritis, ankylosing spondylitis or disc herniation) were excluded before analyses in order to limit the follow-up analysis to those with unspecific symptoms. Hence, of the 1,405 men participating at both surveys, 1,347 constitute the study cohort.
The outcomes were primary care consultations, hospital admissions, sick leave and disability pensions owing to a neck or back diagnosis (International Classifications of Diseases, ICD-9 720–724)  during the follow up period of 12 years. The outcomes were assessed in a structured interview by an experienced physician at survey 2. Each outcome was dichotomized so that one or several consultations, hospital admissions or sick leaves due to neck or back problems was opposed to not having any consultation, admission or sick leave of that kind.
Neck and low back symptoms
All studied variables except rating of psychiatric symptoms were measured at survey 1. Neck pain and low back pain ever during lifetime and during the last year were assessed in a questionnaire . The neck questions posed included neck or shoulder pain/discomfort but this is denoted neck pain in this report to facilitate reading. Three groups were identified: (1) those who reported no neck or low back problems before the study period, (2) those who reported neck or low back problems previously but not during the last year prior to survey 1 (previous neck and/or low back pain) and (3) those who reported neck or low back problems previously and during the year before survey 1 (current neck and/or low back pain). Very few individuals reported neck and/or low back problems only during the last year and not before. These were included in the latter group.
Physical workload was assessed in a structured interview by an experienced physician in occupational medicine as the reported average number of hours working in a sitting or standing position, with a moderate, heavy or very heavy workload during an average working day according to Edholm's activity scale . Educational level was measured on a five grade scale (from compulsory school to university). Experienced demands at work and perceived control in the work situation were assessed in the questionnaire according to Karasek and Theorell . Weight and height were measured with standard procedures and body mass index (BMI) calculated as weight in kilograms divided by height in meters squared (kg/m2). Tobacco and alcohol consumption were assessed in a structured interview. Tobacco consumption was analyzed with two dichotomous variables, namely current daily smoking and current snuff use (smokeless tobacco). Average alcohol intake, computed as grams of pure alcohol consumed per week, was based on frequency of alcohol intake, type of beverage consumed and amount consumed on each occasion. Presence of psychiatric symptoms was assessed with the Hopkins Symptom Check List-25 (HSCL-25) at survey 2 . The scale consists of 25 items covering common psychiatric symptoms in the area of anxiety and depression graded from 1 (not at all) to 4 (very much).
The study was approved by the Research Ethics Committee at the Karolinska Institute in Stockholm, Sweden (Dnr 90:19) and by the Regional Ethical Board, Uppsala, Sweden (Dnr 2005:107). The research was carried out in compliance with the Helsinki declaration and all participants gave their informed consent.
The statistical analyses were conducted using SPSS® version 13.0. Comparisons between groups were done using t-test for continuous data and Chi2-test for categorical data. Multiple logistic regressions, generating odds ratios (OR) with 95% confidence intervals (95% CI), with backward removal of covariates on the 0.10 level were performed to adjust the associations between reported symptoms at survey 1 and outcome measures for possible confounding factors. In the models, educational level was handled numerically as an index ranging from one to five. The internal non-response for various covariates was generally very low. The only exception was demands at work, where 8.8% had missing data. All participants contributed data on outcome measures and BMI. All tests were two-tailed. A p-value of 0.05 was regarded as statistically significant.
In the follow-up cohort based on those 1,347 men who participated in both surveys, 211 men (15.7%) had had at least one primary care consultation owing to neck or back pain, 25 men (1.9%) had been hospitalized for the condition, 73 men (5.4%) had been on sick leave at least once and 39 men (2.9%) had a disability pension owing to a neck or back diagnosis (Table 2). Sixty-four men (4.8%) received a specific back diagnosis during the follow-up period.
The majority, 836 men (62.9%) reported current neck and/or low back pain during the last year prior to survey 1. No neck or low back pain was reported by 269 men (20.2%) and 224 men (16.9%) reported previous neck and/or low back pain. There was some overlap since some men reported only neck pain, some only low back pain and one group reported a combination of both (Table 2).
The four outcomes were clearly overrepresented in the group reporting current neck and/or low back pain at survey 1 (Table 3). Of those 836 reporting current neck and/or low back pain at survey 1, 174 (20.8%) had had at least one primary care consultations due to the condition, 21 (2.5%) had been hospitalized, 61 (7.3%) had been on sick leave and 36 (4.3%) had been granted a disability pension owing to a neck or back diagnosis. Only 14 (5.2%) of those reporting no neck and/or low back pain at survey 1 reported primary care consultations for neck or back pain during follow up, none had been hospitalized, five had been on sick leave and one had a disability pension.
Snuff use, psychiatric symptoms and specific back diagnoses during follow up were positively associated with all four outcomes (table 4). In addition, BMI was positively associated with hospital admission, whereas educational index and control at work were negatively associated with sick leave. Age and smoking were positively associated with disability pension, whereas educational index was negatively related to disability pension.
Current neck and/or low back pain at survey 1 predicted primary care consultations (OR = 4.1, 95% CI 2.2–7.5) and sick leave (OR = 3.2, 95% CI 1.1–9.2) owing to neck or low back diagnoses during follow up, even after possible confounders were taken into account (Table 5). The OR for disability pension was more modified by confounders although the trends were in the same direction as for the other outcomes. Lower age and lower sense of control at work were independent predictors of primary care consultations in the final model, whereas lower education, lower control at work, and more psychiatric symptoms were independent predictors of sick leave. Lower education and snuff use independently predicted disability pensions owing to neck or low back pain diagnoses. Higher age tended to predict disability pension but this did not reach statistical significance. Naturally, specific back diagnoses during follow up were strong independent predictors for all the outcomes. Previous neck and/or low back pain at survey 1 did not predict any of the outcomes before or after multiple adjustments (not shown in tables).
Current neck and/or low back pain at baseline strongly predicted primary care consultations and sick leave owing to neck or low back diagnoses during 12 years of follow up in this cohort of rural middle aged men. However, in general the outcome numbers were relatively low; approximately one fifth of those reporting current neck and/or low back pain at baseline had had a primary care consultation owing to a neck or back diagnosis, and about 7% had been on sick leave and very few had been hospitalized or granted a disability pension owing to the condition. This is in line with findings from other studies showing that neck and low back pain for the majority of patients is a mild symptom/disorder [14, 24]. However, although the vast majority of those reporting neck or low back symptoms do not consult health care providers, several studies indicate that many of these individuals perceive persistent or recurrent symptoms of varying severity [25, 26].
Very few of the men in our study with no current or previous neck or back pain at start reported primary care consultation, hospital admission, sick leave or disability pension owing to neck or back pain during follow up, indicating that few new cases are added in middle age. The fact that relatively few individuals report only previous neck and/or low back pain might depend on two reasons. First, these disorders go with recurrent episodes and therefore those afflicted fall in the "current" group. Second, if one experiences neck and/or low back problems for a shorter period and no recurrent episodes one might be less inclined to remember and report this.
Our study has several strengths. The most important one is the population-based cohort design and the long follow-up time. The participation rate was high considering the efforts required to attend two extensive health surveys 12 years apart. We used validated questionnaires and measurements methods [21–23] when possible. Few methods were available for evaluation of physical workload in large-scale population based studies when this project was designed in the late 1980s. The Edholm scale generates a wide spread of data and large differences between for example farmers and non-farmers  and showed a good correlation with physical work capacity in a sub-maximal work test (unpublished observation). The outcome data was assessed through structured physician-conducted interviews. In a previous analysis we found a good correlation between interview data on hospital admission diagnoses and registered data . Missing data among participators was generally very low. The breadth of the survey with a diversity of variables to consider lowers the risk for confounder bias of our results.
Our study also has limitations. The cohort was rural-based and must be interpreted as such. Significant differences in morbidity and health care utilization have been reported between urban and rural populations, but the impact of this factor is uncertain [28–30]. The cohort included many farmers, a group previously shown to have a relatively low morbidity and to seek less health care in relation to reported symptoms [16, 31]. Therefore, the outcome frequencies found in our study are probably lower than what can be expected in the general population. Another limitation, which might partly account for the low numbers, is recall bias during the 12-year follow up. However, as mentioned we found good congruity between interview information on previous hospital admissions with register data in a previous study . There are no primary care registers available for comparable validation. In addition, this study was restricted to men. Gender differences regarding the studied outcomes might be at hand but this cannot be evaluated from the available data. Neck and low back outcomes were not separated since the conditions were considered to indicate the same type of disorder and, owing to the limited number of cases generating outcomes. There were too few cases with hospital admissions owing to neck or low back diagnoses to allow for analyses. The calculation of ORs for disability pension was also limited by low numbers, rendering a wide confidence interval. A larger sample would probably result in this outcome being statistically significant as well.
The ORs for primary care consultation, sick leave and disability pension during follow up in relation to current symptoms at baseline were only changed to some extent when other covariates were taken into account, even though some of the covariates were independent predictors of the outcomes. Higher education significantly lowered the ORs for sick leave and disability pension, a result that is well in line with the findings of others . Higher rating on the depression and anxiety scale (HSCL-25) was strongly and independently associated with sick leave. However, since the HSCL-25 was assessed at survey 2 this relation is of a cross-sectional nature. Several other studies indicate depression as a strong predictor of disabling neck and low back pain [33–36]. In a parallel study analyzing functional capacity in relation to previous neck or low back symptoms, depression and anxiety was the factor most impressively related to impaired functional physical and social capacity (Thelin and Holmberg, submitted 2006).
Work related factors such as physical workload, farming and demands experienced at work showed no significant relationship with the studied outcomes. It should be noted that the study cohort represents a group with high physical workload as compared with the general population . For the research seeking evidence for causality on physical work related factors and neck and low back pain it is an interesting observation that in this group with high physical workload no significant association between physical workload and subsequent health outcomes could be identified. If heavy physical workload was of major clinical importance for long term health related outcomes due to neck and low back pain this ought to be detectable in our study. It is possible that selection mechanisms may contribute to our finding. However, in a previous study of farmers no healthy worker effect was discerned for musculoskeletal disorders or general morbidity . Among lifestyle factors, smoking, alcohol consumption and BMI held no independent predictive power for any of the outcomes. However, surprisingly, snuff use (smokeless tobacco) strongly predicted disability pension owing to a neck or back diagnosis even after other confounders were considered in the final model. Very few studies exist on snuff use and musculoskeletal symptoms and disorders. A multiple adjusted analysis on infantry conscripts showed snuff use to be an independent risk factor for musculoskeletal injury during a training period . Smokeless tobacco has been proposed as a risk factor for osteoporosis . Our finding, snuff use as a predictor of disability pension owing to neck or back diagnoses, must be interpreted cautiously but certainly warrants further studies, especially as the prevalence of snuff use is increasing in Sweden.
In this study, we have shown that neck and low back symptoms predicted primary care consultations and sick leave owing to neck or low back diagnoses during 12 years of follow up. However, relatively few of the men with symptoms generated health care utilization, sick leave or disability pensions due to the condition. Higher educational level significantly lowered the risk of sick leave and disability pension, whereas depression and anxiety significantly increased the risk of sick leave. These variables seem to be more important to consider than the traditional work related factors focused on in many clinical and rehabilitation settings.
Cassidy JD, Carroll LJ, Cote P: The Saskatchewan health and back pain survey. The prevalence of low back pain and related disability in Saskatchewan adults. Spine. 1998, 23 (17): 1860-1866. 10.1097/00007632-199809010-00012.
Leboeuf-Yde C, Klougart N, Lauritzen T: How common is low back pain in the Nordic population? Data from a recent study on a middle-aged general Danish population and four surveys previously conducted in the Nordic countries. Spine. 1996, 21 (13): 1518-25; discussion 1525-6. 10.1097/00007632-199607010-00005.
Papageorgiou AC, Croft PR, Ferry S, Jayson MI, Silman AJ: Estimating the prevalence of low back pain in the general population. Evidence from the South Manchester Back Pain Survey. Spine. 1995, 20 (17): 1889-1894.
Walker BF: The prevalence of low back pain: a systematic review of the literature from 1966 to 1998. Journal of Spinal Disorders. 2000, 13 (3): 205-217. 10.1097/00002517-200006000-00003.
Alexanderson K, Norlund A: Swedish Council on Technology Assessment in Health Care (SBU). Chapter 1. Aim, background, key concepts, regulations, and current statistics. Scand J Public Health Suppl. 2004, 63: 12-30. 10.1080/14034950410021808.
Ariens GA, van Mechelen W, Bongers PM, Bouter LM, van der Wal G: Physical risk factors for neck pain. Scand J Work Environ Health. 2000, 26 (1): 7-19.
Ekberg K, Bjorkqvist B, Malm P, Bjerre-Kiely B, Karlsson M, Axelson O: Case-control study of risk factors for disease in the neck and shoulder area. Occup Environ Med. 1994, 51 (4): 262-266.
Bernard B: Musculoskeletal disorders (MSDs) and workplace factors. A critical review of epidemiologic evidence for work-related musculoskeletal disorders of the neck, upper extremity, and low back. 1997, Cincinnati , National Institute for Occupational Safety and Health
Burdorf A, Sorock G: Positive and negative evidence of risk factors for back disorders. Scand J Work Environ Health. 1997, 23 (4): 243-256.
Houtman IL, Bongers PM, Smulders PG, Kompier MA: Psychosocial stressors at work and musculoskeletal problems. Scand J Work Environ Health. 1994, 20 (2): 139-145.
Jonsson E, Nachemson A: Back pain, Neck pain. An Evidence Based Review. 2000, Stockholm , SBU - The Swedish Council on Technology Assessment in Health Care, 1:
Linton SJ: A review of psychological risk factors in back and neck pain. Spine. 2000, 25 (9): 1148-1156. 10.1097/00007632-200005010-00017.
Hestbaek L, Leboeuf-Yde C, Manniche C: Low back pain: what is the long-term course? A review of studies of general patient populations. Eur Spine J. 2003, 12 (2): 149-165.
Cassidy JD, Cote P, Carroll LJ, Kristman V: Incidence and course of low back pain episodes in the general population. Spine. 2005, 30 (24): 2817-2823. 10.1097/01.brs.0000190448.69091.53.
Alexanderson K, Hensing G: More and better research needed on sickness absence. Scand J Public Health. 2004, 32 (5): 321-323. 10.1080/14034940410019253.
Holmberg S, Stiernström EL, Thelin A, Svärdsudd K: Musculoskeletal symptoms among farmers and non-farmers: a population-based study. Int J Occup Environ Health. 2002, 8 (4): 339-345.
Stiernström EL, Holmberg S, Thelin A, Svärdsudd K: Reported health status among farmers and nonfarmers in nine rural districts. J Occup Environ Med. 1998, 40 (10): 917-924. 10.1097/00043764-199810000-00013.
Stiernström EL: Cardiovascular morbidity and mortality among farmers and non-farmers. A population-based study. Department of Public Health and Caring Sciences, Family Medicine Section. 2000, Uppsala , Uppsala
Holmberg S, Thelin A, Stiernstrom EL: Relationship of Sense of Coherence to Other Psychosocial Indices. European Journal of Psychological Assessment. 2004, 20 (4): 227-236. 10.1027/1015-57220.127.116.11.
International Classification of Diseases, Ninth Revision (ICD-9), Swedish version. 1986, Stockholm, Sweden , Allmänna förlaget
Ilmarinen J, Knauth P, Klimmer F, Rutenfranz J: The applicability of the Edholm Scale for activity studies in industry. Ergonomics. 1979, 22: 369-379.
Karasek R, Theorell T: Healthy work. Stress, productivity, and the reconstruction of working life. 1990, New York , Basic Books
Nettelbladt P, Hansson L, Stefansson CG, Borgquist L, Nordstrom G: Test characteristics of the Hopkins Symptom Check List-25 (HSCL-25) in Sweden, using the Present State Examination (PSE-9) as a caseness criterion. Soc Psychiatry Psychiatr Epidemiol. 1993, 28 (3): 130-133. 10.1007/BF00801743.
Cote P, Cassidy JD, Carroll L: The treatment of neck and low back pain: who seeks care? who goes where?. Med Care. 2001, 39 (9): 956-967. 10.1097/00005650-200109000-00006.
Croft PR, Macfarlane GJ, Papageorgiou AC, Thomas E, Silman AJ: Outcome of low back pain in general practice: a prospective study. Bmj. 1998, 316 (7141): 1356-1359.
Hestbaek L, Leboeuf-Yde C, Engberg M, Lauritzen T, Bruun NH, Manniche C: The course of low back pain in a general population. Results from a 5-year prospective study. J Manipulative Physiol Ther. 2003, 26 (4): 213-219. 10.1016/S0161-4754(03)00006-X.
Holmberg S, Thelin A, Stiernström EL, Svärdsudd K: The impact of physical work exposure on musculoskeletal symptoms among farmers and rural non-farmers. A population-based study. Ann Agric Environ Med. 2003, 10: 179-184.
Smith MH, Anderson RT, Bradham DD, Longino CFJ: Rural and urban differences in mortality among Americans 55 years and older: analysis of the National Longitudinal Mortality Study. J Rural Health. 1995, 11 (4): 274-285.
Lewis G, Booth M: Are cities bad for your mental health?. Psychol Med. 1994, 24 (4): 913-915.
Watt IS, Franks AJ, Sheldon TA: Health and health care of rural populations in the UK: is it better or worse?. J Epidemiol Community Health. 1994, 48 (1): 16-21.
Thelin A: Morbidity in Swedish farmers, 1978-1983, according to national hospital records. Soc Sci Med. 1991, 32 (3): 305-309. 10.1016/0277-9536(91)90108-O.
Hagen KB, Holte HH, Tambs K, Bjerkedal T: Socioeconomic factors and disability retirement from back pain: a 1983-1993 population-based prospective study in Norway. Spine. 2000, 25 (19): 2480-2487. 10.1097/00007632-200010010-00010.
Carroll LJ, Cassidy JD, Cote P: Depression as a risk factor for onset of an episode of troublesome neck and low back pain. Pain. 2004, 107 (1-2): 134-139. 10.1016/j.pain.2003.10.009.
Croft PR, Papageorgiou AC, Ferry S, Thomas E, Jayson MI, Silman AJ: Psychologic distress and low back pain. Evidence from a prospective study in the general population. Spine. 1995, 20 (24): 2731-2737.
Hurwitz EL, Morgenstern H, Yu F: Cross-sectional and longitudinal associations of low-back pain and related disability with psychological distress among patients enrolled in the UCLA Low-Back Pain Study. J Clin Epidemiol. 2003, 56 (5): 463-471. 10.1016/S0895-4356(03)00010-6.
Jarvik JG, Hollingworth W, Heagerty PJ, Haynor DR, Boyko EJ, Deyo RA: Three-year incidence of low back pain in an initially asymptomatic cohort: clinical and imaging risk factors. Spine. 2005, 30 (13): 1541-8; discussion 1549. 10.1097/01.brs.0000167536.60002.87.
Thelin A, Höglund S: Change of occupation and retirement among Swedish farmers and farm workers in relation to those in other occupations. A study of "elimination" from farming during the period 1970-1988. Soc Sci Med. 1994, 38 (1): 147-151. 10.1016/0277-9536(94)90309-3.
Heir T, Eide G: Injury proneness in infantry conscripts undergoing a physical training programme: smokeless tobacco use, higher age, and low levels of physical fitness are risk factors. Scand J Med Sci Sports. 1997, 7 (5): 304-311.
Spangler JG, Quandt S, Bell RA: Smokeless tobacco and osteoporosis: a new relationship?. Med Hypotheses. 2001, 56 (5): 553-557. 10.1054/mehy.2000.1164.
The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2474/7/66/prepub
This study was financially supported by AFA Insurance, LRF Research Foundation and the Swedish Council for Working Life and Social Research.
The author(s) declare that they have no competing interest.
The two authors contributed equally to the concept of the study, acquisition of data, analysis and interpretation of data and the drafting of the manuscript.