Prevalence and predictors of low back pain in a Southern Nigerian hospital.

Background: Low back pain (LBP) is the most common musculoskeletal disease in adults. The data on LBP from Sub-Saharan Africa are inadequate. The aim of this study was to assess the prevalence and analyze the predictors of LBP among hospital staff in a Nigerian tertiary hospital. Materials and
Methods: The study participants were recruited using an opt-in approach, with the aim of including representative numbers from each professional cadre. Each participant gave formal consent. Ethical clearance was obtained.
Results: Five hundred and sixty-three participants with the mean age of 36.0 ± 8.3 years and 62% female were interviewed. The point prevalence of LBP was 234 (42% [95% confidence interval [CI]: 37%-45%]). Profession was a significant predictor of LBP (P = 0.001) - nurses (53% [95% CI 43%-63%]), administrative officers (49% [95% CI 40%-59%]), engineers (50% [95% CI 24%-76%]), and health information staff (50% [95% CI 26%-75%]) had the highest prevalence. In univariate regression, female gender, increasing age, body mass index ≥25 kg/m2, and frequently adopting a bending posture, were significantly associated with LBP, while in multivariate regression, only the female gender was a significant predictor.
Conclusion: The pattern of both the professions at risk, due to the well-known mechanisms of poor ergonomics, and the marked risk for the female gender, in the hospital setting, suggest underresourced work and societal environments as the underlying factors-more research is needed.

INTRODUCTION

Low back pain (LBP) is the most common musculoskeletal complaint in adults, affecting 70% in high-income countries at least once in a lifetime.[1] Its prevalence ranges from 26% to 79% in high-income countries and 16%–59% among adults in Sub-Saharan Africa, where it is generally assumed that the prevalence of LBP is lower, probably due to a much smaller number of studies done on the continent.[2] In addition, its worldwide impact is enormous – 100 billion dollars lost by the US economy annually (direct and indirect), more than 700 Euros per person in Germany (direct), and in the 2010 Global Burden of Disease Study, it was the seventh cause of Disability-Adjusted Life Years (ahead of road traffic injury).[345]

The data from Sub-Saharan Africa on musculoskeletal disease are inadequate, in part due to poor health research financing, priority given to epidemics such as HIV/AIDS and an underdeveloped research culture.[2] What data are available is often derived from the experience of people working in a particular setting or from patients, probably because this may require fewer resources than a study in the general population.[26] The Nigerian economic landscape is that of a low-to-middle income country, where there is low-grade system support for workers – lack of stable electricity and social support services.[7] Consequently, the workers labor often without attention to ergonomics and with scanty provision of assistive devices to facilitate lifting of loads such as patients.[8] However, like other countries, women are entering the workplace in large numbers, albeit with cultural expectations of their subservient role in a hierarchical society, intact.[8]

LBP is defined as pain, muscle tension, or stiffness below the 12th rib and above the inferior gluteal folds, with or without leg pain, it becomes chronic when it lasts longer than 3 months, lasts longer than time expected for healing to have occurred, or recurs frequently.[19] The etiology of LBP can be primary, secondary to systemic disease and nonorganic. Primary LBP can be mechanical or non-mechanical, while non-organic LBP is strongly associated with psychosocial experience and occupational LBP.[10] While acute LBP often has defined pathology and is expected to resolve most of the time, chronicity is poorly correlated with definite pathology and tends to respond inadequately to treatment.[10] Rest and analgesia were noted as the most common modality of treating LBP in Africa, despite evidence indicating “active rehabilitation” involving “guided physical exercise” as most likely to reduce recurrence or curtail chronicity.[210]

Previous large (>400 participants) hospital-based studies in Nigeria have focused on nurses and thus were constrained in adequately addressing the prevalence and problems of LBP in other hospital workers. Therefore, we aimed to undertake a study of LBP in all cadres of staff in a fairly large Nigerian hospital, to obtain data on the demography and predictors of occupational LBP in a low-to-middle income setting.

MATERIALS AND METHODS

Study design

This study was carried out in the University of Uyo Teaching Hospital, a tertiary (referral) hospital in southern Nigeria. We undertook an institution-based cross-sectional study, using a self-administered previously published structured questionnaire – some aspects, such as anthropometric measurements, were interviewer assessed.[11] All participants signed written informed consent.

The questionnaire has 29 items and incorporates the Oswestry Disability Index Questionnaire in an optional section for respondents with LBP.[12] Other sections in the instrument are sociodemographic characteristics, Stevenson's family staging, LBP, past medical history, and anthropometric measurements.

The sample size was calculated using Yamane's formula:

n = N/1 + N (e)2

Where n = sample size; N = population (total number of staff-2050); e = 1 – precision; precision = 95% confidence interval (CI).[13]

The calculated sample size was 334.

Source population

The participants were recruited by approaching all staff at their duty posts in different sections of the hospital over the period of the study, from January to March 2018. The staff were given the option to participate in the study and those who chose to do so gave formal consent before filling the questionnaire. Any areas that needed clarification were explained and anthropometric measurements were carried out by the researchers and their assistants. High-pressure zones such as accident and emergency and labor ward were also included in recruitment of participants, with adequate precautions. We excluded obviously pregnant women and recent victims of major accidents resulting in back injury.

Data analysis

This was carried out with the use of XLSTAT add-in to Microsoft Excel supplemented with STATA 15.1 (StataCorp, Texas, USA) for regression analyses. The Marascuilo procedure was used to simultaneously test for statistical difference for all pairs of proportions. Univariable and multivariable logistic regression models were built to determine factors independently associated with LBP. All variables with P value of the Wald statistic of <0.25 at the univariable level were included in the multivariable model. Model diagnostics were performed using a receiver operator characteristic curve.

Ethical considerations

Ethical approval was obtained from the Institutional Health Research Ethical Committee.

RESULTS

Five hundred and sixty-three hospital workers with the mean age of 36.0 ± 8.3 years participated in the study. Most respondents were female – 347 (62%). Table 1 summarizes the sociodemographic and anthropometric characteristics of the participants. The distribution of the various professions among the responders is shown in Figure 1 – doctors have the highest response – 148 (26%), followed by administrative officers – 113 (20%) and nurses – 100 (18%). The median duration of employment at this hospital was 6 years (interquartile range: 1–10 years). The response rate for the Oswestry Disability Index outcome measure was 89% (208) of the participants with LBP.

Table 1

Sociodemographic and anthropometric characteristics of the participants

Variable	Value, n (%)	P
Age	36.0±8.3	0.002*
Groups: 20-39 years; 40-60 years
Mean
Gender
Female	347 (62)	<0.0001†
Male	216 (38)
Educational level
Primary	4 (1)	0.165‡
Secondary	41 (7)
Tertiary	518 (92)
Income level naira (USD)
<50,000 (140)	97 (17)	0.013§
51-100,000 (141-280)	135 (24)
101-150,000 (281-420)	123 (22)
151-200,000 (421-560)	61 (11)
>200,000 (560)	147 (26)
Alcohol use	164 (29)
Tobacco use	5 (1)
Exercise
Regular exercise	223 (40)	0.373‖
Median exercise per week	3 (IQR: 2-7)
Weight (kg) - mean	73.0±13.5
Height (m) - mean	1.64±0.09
BMI (kg/m2)
Groups <25 kg/m2; ≥ 25 kg/m2	27.2±4.9	0.013¶
Mean

*The difference between the proportions of the two age groups is significant - [Table 5], †P<0.0001 - [Table 2], ‡Not significant, §P - 0.013 no significant difference between the income levels, ‖Not significant. ¶P - 0.013 - [Table 5]. BMI=Body mass index

Figure 1

A column chart showing the response rates of the various cadres of staff. *Others – miscellaneous group including 2 drivers, 1 builder, 3 carpenters, 6 auditors etc

Table 5

Prevalence of low back pain among age groups 20-39 years and 40-60 years; body mass index groups <25 kg/m2 and ≥25 kg/m2

	LBP (n=234; 42%), n (%)	No LBP (n=329; 58%), n (%0	Total (n=563), n (%)	P
Age (years)
20-39	144 (37)	242 (63)	386 (69)	0.002*
40-60	90 (51)	87 (49)	177 (31)
BMI (kg/m2)
<25	70 (35)	132 (65)	202 (36)	0.013†
≥25	164 (45)	197 (55)	361 (64)

*P - 0.002 the difference between the proportions of the two age groups is significant, †P - 0.013 the difference between the two BMI groups is significant. BMI=Body mass index, LBP=Low back pain

Gender was a significant predictor of LBP (P < 0.0001) [Table 2]. Table 2 also shows the distribution of disability from LBP among the participants, with respect to gender. The difference in proportions of female and male with LBP (48% and 31%, respectively) was significant, indicating the female gender as being at risk for LBP. Logistic regression showed that at the univariate level, females were two times more likely to have LBP [Table 3].

Table 2

Prevalence and disability level of low back pain in the female and male gender

Gender	LBP (n=234; 42%), n (%)	No LBP (n=329; 58%), n (%)	P	Total (n=563), n (%)	Disability level (n=208; 89%)
					Minimal (n=166)	Moderate (n=39)	Severe (n=2)	Crippled (n=1)
Female	168 (48)	179 (52)	<0.0001	347 (62)	118 (71)	33 (85)	2 (100)	1 (100)
Male	66 (31)	150 (69)	-	216 (38)	48 (29)	6 (15)	0	0

The difference between the proportions of female and male gender is significant. LBP=Low back pain

Table 3

Univariable and multivariable logistic regression models

	Univariable		Multivariable*
	OR (95% CI)	P	OR (95% CI)	P
Age	1.03 (1.01-1.05)	0.004	1.02 (0.99-1.04)	0.19
Gender
Male	1		1	<0.001
Female	2.08 (1.46-2.98)	<0.001	1.93 (1.33-2.79)
Predominant posture
Forward bending	1		1
Sitting	0.44 (0.19-1.01)	0.05	0.48 (0.20-1.14)	0.10
Standing	0.53 (0.23-1.18)	0.12	0.63 (0.27-1.45)	0.28
BMI (kg/m2)	1.02 (1.00-1.05)		0.98 (0.93-1.02)	0.29

*Area under the ROC curve=0.63. OR=Odds ratio, CI=Confidence interval, BMI=Body mass index, ROC=Receiver operator characteristic

The total number of respondents with LBP at the time of the study (point prevalence) was 234 (42% [95% CI 37%–45%]). For 76 (33%) of them, this was the first episode; 49 (65%) of these first timers were in the 20–39 years age group. Profession was a significant predictor of LBP (P = 0.001) – nurse (53% [95% CI: 43%–63%]), administrative officers (49% [95% CI: 40%–59%]), engineers (50% [95% CI: 24%–76%]), and health information staff (50% [95% CI: 26%–75%]) had the highest prevalence [Table 4]. The significant differences were between the LBP prevalence of nurses and administrative staff, respectively, and pharmacists.

Table 4

Prevalence of low back pain in the various professions of the study participants

Demographic characteristics	Prevalence of LBP				Total (n=563)
	LBP (n=234; 42%)		No LBP (n=329; 58%)
Occupation	n (%)	95% CI	n (%)	95% CI	n (%)	95% CI
n>40
Nurse	53 (53)	43.2-62.8	47 (47)	37.2-56.8	100 (17.8)	14.6-20.9
Administrative officer	56 (49)	40.3-58.8	57 (50)	41.2-59.7	113 (20.1)	16.8-23.4
Others*	23 (48)	34.6-63.2	24 (51)	36.8-65.4	47 (8.3)	6.1-10.6
Health assistant	19 (39)	25.8-53.4	29 (60)	46.6-74.3	48 (8.5)	6.2-10.8
Medical doctor	45 (30)	22.9-37.8	103 (69)	62.1-77.0	148 (26.3)	22.7-29.9
n<40
Engineer	7 (50)	23.8-76.2	7 (50)	23.8-76.2	14 (2.5)	1.2-3.8
Health information officer	8 (50)	25.5-74.5	8 (50)	25.5-74.5	16 (2.8)	1.4-4.2
Physiotherapist	4 (44)	11.9-76.9	5 (55)	23.1-88.0	9 (1.6)	0.5-2.6
Medical laboratory scientist	11 (44)	24.5-63.5	14 (56)	36.5-75.5	25 (4.4)	2.7-6.1
Optometrist	4 (40)	9.6-70.4	6 (60)	29.6-90.4	10 (1.8)	0.6-2.9
Pharmacist	4 (12)	0.9-23.3	29 (88)	76.7-99.0	33 (5.9)	3.9-7.8

P - 0.001. Significant difference noted between proportions. 1. Nurse and pharmacist, 2. Administrative officer and pharmacist. *Others - miscellaneous group including 2 drivers, 1 builder, 3 carpenters, 6 auditors etc., BMI=Body mass index, LBP=Low back pain

There was no statistical difference between the various standard strata of the body mass index (BMI) and the occurrence of LBP – P = 0.07. Further analysis revealed a BMI of 25 kg/m2 as a threshold above which there was a significant association with LBP P = 0.013 [Table 5]. It was noted that 193 (35%) had a normal BMI, while 361 (64%) were overweight. At the univariate level, for every 1-year increase in age, there was a 3% increase in the likelihood of having LBP [Table 3]. P value for the age series shown in Table 5 was 0.002, indicating that there was a significant association between age and LBP. The Marascuilo procedure revealed the age group of 40–60 years was a significant risk factor for LBP. Multiple logistic regression models revealed the female gender as the only significant predictor of LBP when gender, age, BMI, and posture commonly adopted were considered (P < 0.001) [Table 3].

Marital status with respect to the risk of developing LBP had P = 0.009 indicating a significant relationship, but the Marascuilo procedure did not yield information on which specific status had a difference large enough to be significant. The data on posture were obtained in a form that requests the study participants indicate the posture they commonly adopt. In this framework, P value for LBP as related to bending was 0.001, with the Marascuilo procedure confirming a difference large enough to be significant. The other postures, standing and sitting, were not significantly related with LBP (P > 0.05). The relation of mattress on which the participant sleeps (soft, firm, orthopedic, none – bare floor or mat) to LBP was significant (P = 0.002), but there was no significant difference between the various types of mattress but rather between any mattress type and sleeping without a mattress.

DISCUSSION

The point prevalence of LBP found in this study was 42% (95% CI: 37.1%–45.4%).

The professions with the highest prevalence of LBP were nurses, administrators, engineers, and health information. Nurses and administrative staff were found to be significantly at risk for LBP. In addition, female gender, increasing age, BMI ≥25 kg/m2, and frequently adopting a bending posture were significantly associated with LBP.

Some strengths of this study are the large sample size, and the use of point rather than 12-month prevalence to reduce recall bias. Details of the possible weight of patients carried by staff were not feasible because the data were not available. Our questionnaire did not include any items on psychosocial aspects. Study participants were unintentionally skewed toward the well-educated, suggesting some sampling bias. However, background knowledge indicates that for most hospital professions, university education is a minimum. This study does not include data on participants above 60 years because that is the retirement age. Representative recruitment of the various professions was achieved, the total number of participants being 25% of the staff population. Applying minimal exclusion criteria increased the external validity of the results, enhancing generalizability.[14]

Omokhodion et al. (Nigeria) had a smaller sample size but obtained similar results with respect to the prevalence of LBP (46%), professions with the first and second highest prevalence of LBP being nurses and administrative staff respectively and with the female gender being at increased risk of LBP.[15] Their study did not highlight the problem of LBP in minority professions. We found that engineers and health information staff had a high prevalence of LBP. Karahan et al. (Turkey) also noted the profession with the highest prevalence of LBP to be nurses, but they found a much higher prevalence among physicians and physiotherapists than in the present study (63% and 73%, respectively).[16] This difference from our study may be due to the recruitment of large number of staff in these professions (total sample size of 1600) and the application of 12-month prevalence, which should be higher than point prevalence and may be subject to recall bias.[14] The mechanism of increased LBP in the nursing profession is thought to be faulty ergonomics – heavy lifting with limited lifting aids, bending, twisting, and static postures.[17] A similar mechanism could apply among engineers, though there is a paucity of studies on this subgroup in the hospital setting.[18] The work profile of administrative and health records staff is similar, long sitting while working either at computers or sorting medical records. This has been noted as a major risk factor for occupational LBP, especially in relation to inefficient ergonomics in the workplace, predisposing to repetitive stress injury.[19]

Violante et al. (Italy) carried out a study on LBP among female nurses, which revealed differences in the outcome with respect to the type of LBP specified.[20] The overall prevalence of LBP was 44%, which is similar to that in our study, but unlike our results, age was a significant risk factor for LBP only among the lumbar disc hernia subgroup, while BMI was not a risk factor. Our findings that age above 40 years was significantly associated with LBP may correlate with their finding about its relationship with lumbar disc hernia, which is more common above 40 years of age. In addition, we noted a BMI threshold of 25 kg/m2 above which there was a significant risk for LBP, a finding also noted by Spyropoulos et al. (Greece) and Deksisa Abebe (Ethiopia).[2122] While the findings by Spyropoulos were similar to ours, Abebe found that age and gender were not risk factors for LBP. Their findings in the age category may have been due to the majority of their respondents being below 40 years (77%), going by the 40-year threshold in our study, above which we found age to be a risk factor.

The female gender is a notable risk factor for LBP in Sub-Saharan Africa and Europe.[221] Our finding that the female gender is twice at risk of LBP and the only factor to remain significant after multivariate regression, emphasizes the importance of this risk factor. Meucci et al. noted the possible etiologies to include musculoskeletal (pregnancy, childcare, and double workday), physiological (less muscle and bone mass), and psychological factors.[23] The underresourced and patriarchal environment in Sub-Saharan Africa, which could exacerbate the musculoskeletal and psychological factors, requires evaluation by further studies and consideration by policy-makers to reduce the prevalence of LBP among females in the work place.

The posture which was significantly associated with LBP was bending (P = 0.001). Bejia et al. (Tunisia) found no relationship of LBP with posture, while Spyropoulos et al. noted that only the lifetime prevalence of LBP was associated with being bent forward while sitting for more than 2 h, while sitting with or without a back support had no relationship with point, one or 2 year prevalence.[2124]

The relation of LBP to the type of mattress was studied by Kovacs et al. using a randomized double-blind controlled design.[25] They found medium-firmness mattresses to be associated with reduced LBP compared with firm mattresses. Our finding that sleeping without a mattress is associated with increased LBP is similar because a floor or mat is much harder than a firm mattress, though our study found that the type of mattress is not a risk factor. This difference in our findings could be accounted for by our not determining the mattress type (we used the patients' perceptions), which Kovacs et al. accounted for by supplying the mattresses, having graded each.

Nurses and administrative staff are most at risk for LBP in the hospital setting. However, attention should be paid to minority professions, health information staff and engineers. The female gender is markedly at risk for LBP, and further research is needed to define causation. Policy is also needed to reduce the incidence of LBP in working women, especially in underresourced settings.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.