Criteria for diagnosis and attribution of an occupational musculoskeletal disease.

BACKGROUND: Criteria for diagnosis and compensation of occupational musculoskeletal diseases varies widely between countries as demonstrated by the large differences between countries with comparable economics and social systems (for example, within the European Union). Several countries have a list of occupational diseases and sometimes these lists include diagnostic and attribution criteria, but these criteria are usually not very specific, and they may also be very different.
OBJECTIVES: The aim of this paper is to explicitly define what are the information needed for an evidence-based diagnosis and attribution of an occupational musculoskeletal disease.
METHODS: Based on the general framework of evidence-based medicine, a review is presented of the information required to define: - when a musculoskeletal disease is present, according to the best available techniques; - how to define a relevant exposure to biomechanical risk factors, according to the best available techniques.
RESULTS: Criteria are presented to combine information regarding the diagnosis of a musculoskeletal disease and exposure to biomechanical risk factors for an evidence-based attribution of the disease to the occupational exposure. The criteria use a probabilistic model that combine epidemiologic and medical findings, workplace exposure assessment, and non-occupational factors evaluation. DISCUSSION: The use of the proposed criteria may improve the process of diagnosis and attribution of an occupational musculoskeletal disease. In addition, it makes possible to associate a probability rank to the attribution and, ultimately, it may improve the overall quality of the decisional process of the occupational physician.

Background

Occupational musculoskeletal diseases (MSDs) have a significant impact on worker experienced pain and disability, health care costs and lost time from work. For example, most episodes of low back pain (LBP) do not have an identified cause, yet a sizable part of them are considered to have been induced by work activity. It has been estimated that in 2017, world-wide, low back pain affected more than half a billion of persons and that it caused almost 65 million of Years Lived with Disability (8). World-wide, it has been estimated that in 2017 low back pain induced by the exposure to “occupational ergonomic factors” was responsible for 25% of all Disability Adjusted Life Years (DALYs, Years of Life Lost plus Years Lived with Disability) attributable to occupational risks (9).

The determination whether an MSD is “occupational” is an important issue in many settings, including for physicians treating individual patients and considering if work exposures might be causing or exacerbating the condition, for workers to seek recognition of the disease under workers’ compensation statutes, for public health surveillance efforts and epidemiological studies, and for controls of workplace exposures. Lack of a common framework for defining whether an MSD is “occupational” impairs all these efforts. Differences in definitions lead to large variations in national statistics and other reporting of occupational diseases and may ultimately affect workers and employers in several ways.

Before diving into the topic, however, I would like to preliminarily define two terms (disease and occupational) which are purposefully used here (and the consequent choice of avoiding the use of disorder and work-related).

The first, is the use of disease and the choice of not using disorder. In the literature concerning MSDs it is common to see the use of “disorder” interchangeably with “disease”, even if the two terms point to different conditions.

Disease denotes a condition characterized by functional impairment, structural change, and the presence of specific signs and symptoms. Disorder, in contrast, denotes a condition characterized by functional impairment without structural change and, while certain disorders or categories of disorders might be accompanied by specific signs and symptoms, their presence is not required for a condition to be termed a disorder (4). Hence, the presence of pain in a body region does not necessarily indicate the presence of an underlying pathology, especially in the case of musculoskeletal pain. In addition, when musculoskeletal pain is present, the findings of a reduction in the range of motion should be taken as granted (antalgic defense mechanism) and the condition is still compatible with a disorder. Because this paper deals with occupational diseases, no discussion will be made of the common condition of musculoskeletal pain, that is, pain without any definite structural change of the body even if it is known that pain (usually transient) in a body region may be due to occupational activity, as well as other physical activity.

The second, is the use of occupational and the choice of not using work-related. In the literature concerning MSDs it is common to see the use of “work-related” interchangeably with “occupational”, even if the two terms point to different conditions. According to the original definition of the WHO working group (6) “… in occupational diseases, there is a direct cause-and-effect relationship between hazard and disease. In work-related diseases, in contrast, the work environment and the performance of work contribute significantly, but as one of a number of factors, to the causation of a multifactorial disease. Occupational diseases therefore stand at one end of the spectrum of work-relatedness, where the relationship to specific causative factors at work has been fully established and the factors concerned can be identified, measured, and eventually controlled. At the other end, diseases may have a weak, inconsistent, unclear relationship to working conditions; in the middle of the spectrum there is a possible causal relationship but the strength and magnitude of it may vary.”

In the text of this paper, “occupational” is to be intended as “caused by work”. This includes the case of a disease for which:

the work activity is an exclusive cause, or

the work activity is a necessary contributory cause, together with other factors (i.e. the disease would not have occurred, at that time, if the person had not performed that work activity), or

the work activity has influenced the course of the disease, which would have been more favourable if the person had not performed that work activity (a case which is usually referred to as a “work-exacerbated disease”, as, for example, in the case of asthma).

For these reasons, only the term occupational is used in this manuscript.

Many countries and some international organizations (e.g., International Labour Organization -ILO) have developed their own lists of occupational diseases (the so called “scheduled diseases”): these lists may specify minimum medical diagnostic and workplace exposure criteria needed for a worker to receive social insurance benefits for a particular disease. Most such lists include well-known occupational diseases, such as noise induced hearing loss and asbestos induced mesothelioma: however, there is a great deal of variability across countries, and common MSDs (e.g., carpal tunnel syndrome) are absent from many lists even though they may be due to occupation.

MSDs may be poorly represented on many countries’ lists of occupational diseases for several reasons. For most MSDs there is no “gold standard” case definition and often no non-invasive investigation that can conclusively confirm the diagnosis. Biopsies to confirm the diagnosis are difficult to obtain, or even unethical, given the risk-to-benefit ratio of some procedures. Furthermore, the causes of MSDs are thought to be multi-factorial: although some MSDs may be associated with workplace exposures, their most prevalent causes in the general population are related to other factors such as age, obesity, sex, and some co-morbidities.

While it may be difficult to establish criteria for occupational MSDs, the effort is important in order to provide compensation to injured workers and to encourage employers and governments to prevent these diseases. In addition, common criteria can improve medical management of injured workers through evidence-based treatments and more accurate prognoses. Clear criteria will improve the assessment of the real “burden of disease” linked to specific risk factors, which is necessary for planning preventive interventions and the adoption of clear criteria will also address the twin problems of under- and over-reporting of occupational MSDs (7). Under- and over-reporting can be strongly influenced by non-medical factors, such as financial incentive to the worker, payments to health care providers, to attorneys or consultants, derived costs to the company and so on.

This paper proposes a methodological framework for developing criteria for diagnosis and attribution of occupational MSDs: the approach follows evidence-based medicine criteria and medical best practices (11).

Methods

Occupational diseases are fundamentally defined by their etiologic characteristics: therefore, linking a disease to an occupational exposure requires several logical steps. The first step is making an appropriate diagnosis of the clinical condition; the second step is assessing the extent of the workplace exposure and the third step is determining if workplace exposure is likely to have caused or exacerbated the condition. This latter step involves as a prerequisite a thorough assessment of the existing medical literature on the association between exposure and disease.

The diagnosis of a potentially occupational MSD follows the typical procedures of clinical medicine with history taking, a focused physical examination, and relevant laboratory and/or imaging and/or pathology studies. The finding from each step of this process can increase or decrease the confidence that the disease is truly present. For example, the presence in a patient of numbness in the thumb, index and middle finger with a positive Phalen’s test is associated with the possible presence of carpal tunnel syndrome (16). However, if a nerve conduction study confirms the presence of median nerve slowing, the likelihood that the carpal tunnel syndrome is present increases to probable.

Beyond this first step of making a clinical diagnosis, additional steps are necessary to identify whether there is a causal link to a workplace exposure, based on the analysis of workplace risk factors. Only very few diseases rarely occur without an occupational or environmental exposure (for example most males with mesothelioma have had asbestos exposure). Additionally, most diseases caused or exacerbated by work are clinically identical to the same diseases that do not have an occupational cause. Lateral elbow tendinopathy occurring in a sedentary worker who recently started playing tennis is clinically indistinguishable from the same disease occurring in a person who recently started a job that required continuous forceful gripping and wrist extension. Allergic dermatitis resulting from a workplace exposure may have the same clinical appearance as allergic dermatitis resulting from an exposure outside the workplace. Documentation of a substantial exposure increase the likelihood that a given disease in a given person is due to work activity.

For causal attribution of MSDs, it is necessary to estimate exposures to biomechanical factors in the workplace, through history taking and review of all available data, to confirm the qualitative, quantitative and temporal parameters of the risk factors and their associations to the clinical history. Like the confirmation of disease, the causal contribution from the exposure can also be considered on a probabilistic scale.

A final step is the review of the scientific literature addressing the associations of the occupational risk factors to the disease. For an MSD to be considered occupational, it should be determined whether the workplace biomechanical exposures are adequate in intensity and overall duration to injure the musculoskeletal tissues in question. This may translate into the following questions:

From the collection of available data, are potentially causal exposures identifiable?

Is there an exposure factor (or a set thereof) in the work activity to which the worker has been exposed for the minimum exposure time and at the minimum intensity needed to induce the disease?

Is the latency of the disease appropriate based on the history?

Has the role of non-occupational risk factors been considered?

As noted above, the recognition of an occupational disease should consider three conditions (5, 10, 20):

(i) the diagnosis, confirmed according to the best applicable criteria for a specific “disease”;

(ii) the assessment of workplace exposure to a given risk factor, performed in accordance with the best applicable criteria, specifying quantitatively frequency (how many times takes place), duration (how much time, altogether, in the working life) and intensity;

(iii) the determination, based on the literature, that the exposure is sufficient to cause the disease.

The proposed criteria have the following elements:

criteria for the quality of the diagnosis of an MSD (diagnostic criteria);

criteria for the quality of the assessment of an exposure to occupational biomechanical risk factors;

criteria for the overall likelihood that a worker has an occupational disease (including criteria for the likelihood that non-occupational factors are the cause of the disease).

Results

Criteria for the quality of the diagnosis of an MSD (diagnostic criteria)

Criteria for the MSD should consider the existing clinical classification based on multidisciplinary guidelines developed by scientific societies. Diagnostic criteria should include relevant symptoms, clinical examination findings, and other tests as appropriate, such as imaging or nerve conduction studies.

Criteria for diagnosis may be assigned a probability of presence of the disease based on progressive levels of evidence (e.g., not likely, possible, probable, very probable, near certain). In this context:

“insufficient or not likely” may be conceptualized as a very small probability (e.g., 0 to 10%),

“possible” may be conceptualized as a probability in the range 10-40%,

“probable” may be conceptualized as a probability in the range 40-60% (that is, the probability may be in favour of the disease, but also against it, so clinical judgment is required),

“very probable” may be conceptualized as a probability in the range 60-90%,

“near certain” may be conceptualized as a very large probability (e.g., 90 to 100%).

The assignment of probability should be based on a review of the literature guided by well-accepted evidence-based approaches for evaluating the literature, such as, for example, the GRADE system (2, 14, 19, 22). The GRADE system takes into account the different issues which may be related to bias in both observational and randomized studies.

Diseases in the same body region, may exist along a spectrum that should be acknowledged. For example, in the low back region, the condition may be non-specific LBP; LBP with sciatica; LBP with clinically significant disc narrowing; or LBP with clinically significant disk protrusion (extrusion). When possible, the discussion should consider whether these are different diseases or different presentations of same disease.

Studies of MSDs should use a minimum criterion for classifying cases based on a set of symptoms and physical examination findings and a reference test, assumed as “gold standard”, such as an imaging study. It is unusual for published studies to have evidence of an MSD based on pathological specimens or biopsies (18). If there is no reference test, all the available evidence including effects of therapeutic interventions must be evaluated (15).

Symptoms should be distinguished from the findings of the clinical examination. If clinical examination manoeuvres are an important element in diagnosis, it should be born in mind that small differences in how the manoeuvres are conducted may produce very different findings (12).

A peculiar issue linked to the diagnosis of MSD is the real objective character of the physical examination. Conventional medical wisdom assumes that the physical examination is an objective assessment: it is, in reality, but not always. Hearth sound assessed by a doctor are objective, in that they do not depend on the cooperation of the patient (and there are examples of physical examination findings, for example, in unconscious patients). With reference to the physical examination of MSDs patients, the situation is different in that, if a few findings are really objective (e.g., circumference of an arm for evaluation of muscle wasting, or passive range of motion ignoring possible induced pain), most of the physical manoeuvres recommended for the diagnosis of MSD are in effect pain inducing manoeuvres, that is, they rely upon the subjective response of the examined person (e.g., Tinel’s or Lasegue’s signs). As physical examination has been defined as “the process of evaluating objective anatomic findings through the use of observation, palpation, percussion, and auscultation” (23), a manoeuvre which rely upon the response of a subject may not be defined as really objective.

Given these considerations, an example of different criteria for the quality (and the associated likelihood) of the diagnosis of tendinopathy of the shoulder is given in Table 1.

Table 1

Different criteria for the quality (and the associated likelihood) of the diagnosis of tendinopathy of the shoulder

Diagnostic criteria	Level of evidence
Neck pain with palpation tenderness over the trapezius muscle.	Insufficient (may just be transient discomfort)
Shoulder pain worsened by abducting the upper arm.	Possible
Shoulder pain with clinical findings of palpation tenderness at the acromion and pain with active arm abduction greater than 45 degrees (13).	Probable
Symptoms, clinical findings and magnetic resonance imaging study that reports well described alterations of the tendon signal in the affected side and (if it is the case) absence of alterations in the asymptomatic shoulder.	Very Probable
Report of surgical intervention with documentation of tendon degeneration or laceration and histology demonstrating tendinosis.	Near Certain

Levels of evidence are critical in order to correctly frame the interpretation of the diagnosis since the minimum level of evidence is likely to vary depending on the actions to be taken (see the preceding table). For example, the level of evidence may be different for the collection of national statistics than for the determination of financial compensation. For insurance or civil litigation, most systems require a minimum legal criterion of “more likely than not”, which would correspond to a level in the range of “very probable” (or higher) in the preceding table, whereas for a criminal case a fact should be assessed “beyond any reasonable doubt”, which would correspond to a level of “Near Certain”. The level of evidence is also likely to influence recommended treatments: for example, an injection of corticosteroid into the shoulder joint may be recommended with a level of “Possible” while surgery may be recommended with a level of “Very Probable”.

Criteria for the quality of the assessment of an exposure to occupational biomechanical risk factors

The assessment of the workplace exposure to the presumptive biomechanical risk factors should be defined both qualitatively (ensure that the subject has actually been exposed to the factor) and quantitatively (ascertain the overall amount of exposure: e.g., level, frequency and cumulative duration).

The evaluation of the exposure can be articulated for progressive levels of evidence both from the qualitative point of view (roughly: possible, probable, very probable, near certain) and from the quantitative point of view (by measures obtained by validated methods). If continuous measures are not available (e.g., total hours of exposure to a vibrating tool with handle vibration acceleration levels), it will be necessary to define the exposure in a semi-quantitative way, with reference to the levels that have been associated by the scientific literature with a significant risk of disease (for example, for knee osteoarthritis see 22).

Given these considerations, an example of different criteria for the quality (and the associated likelihood) of exposure to biomechanical risk factors and the probabilistic strength of the evidence of the exposure (irrespective of the intensity) is given in Table 2.

Table 2

Different criteria for the quality (and the associated likelihood) of exposure to biomechanical risk factors and the probabilistic strength of the evidence of the exposure (irrespective of the intensity)

Exposure Assessment	Level of evidence
Exposure described by the worker or assumed based on a job title.	Insufficient (exposure to relevant biomechanical risk factors may not be objectively assessed based only on a subjective description or on a job title)
Administrative documentation of employment at a company, job title and some written documentation about the work from which it is possible to assume the exposure, at least qualitatively.	Possible
Written evaluation by a safety professional who has observed the job and documented biomechanical exposures by means of an appropriate checklist or similar tool.	Probable
Video analysis with some measurements of risk factors at the job site (e.g., weight of tool or parts, estimation of postures, duty cycles, and so on).	Very Probable
Measurement of movements of the body or body parts of the worker by means of detailed video recordings and/or inertial measurement units, measurements of force exerted by means of mechanical sensors during the execution of the task and so on.	Near certain

Criteria for the overall likelihood that a worker has an occupational disease

Once the level of quality/probability of the disease diagnosis and the quality/probability of exposure are known, the occupational physician must combine this information into a statement about the relation of the disease to the occupation, taking into account all the available evidence, including the non-occupational risk factors present in a given worker, using information deriving by evidence-based review of the relevant scientific literature.

An evidence-based review of the literature evaluating the relationship of workplace factors to disease should follow high quality methods (1, 2, 14, 19). The review may include meta-analyses but should evaluate study quality issues such as bias and power. Studies may evaluate the risks associated with specific industries or jobs: however, when available, the relationships of specific workplace factors or risk models should be quantified with dose-effect relationships.

Table 3 presents a scheme for classifying the quality of studies investigating the relationship between exposure to biomechanical risk factors and musculoskeletal diseases.

Table 3

Scheme for classifying the quality of studies investigating the relationship between exposure to biomechanical risk factors and musculoskeletal diseases

			Exposure assessment
			Objective evaluation		Indirect evaluation
			Quantitative method of direct measurement	Video analysis or video-based observations or experimental protocols	Experts’ observations	Job title, self-reported assessment, job exposure matrix
Case definition	Objective diagnostic criteria	Imaging (plus physical examination)	++/++	++/+	++/-	++/--
		Physical examination (symptoms plus clinical signs)	+/++	+/+	+/-	+/--
	Symptoms	Structured interview (current and past health history)	-/++	-/+	-/-	-/--
		Self-administered questionnaire	--/++	--/+	--/-	--/--

It is quite clear that only studies that can be classified in the box (++ / ++) are truly informative in that they investigate, in the correct way, both exposure and disease. Unfortunately, these studies in the scientific literature are extremely rare, while studies that have used questionnaires for both the definition of the exposure and the disease abound, without being informative, as has been authoritatively noted, for example, in the field of studies on low back pain: “One of the pitfalls in back pain research is that both back pain and its exposures can be simply assessed by a questionnaire. However, most of these measurements provide reliable and valid estimates of neither exposure nor outcome. In our view, this is one of the most important reasons for the continuing controversy about the causes of back pain, with some authors negating an association and others advocating important associations with work.” (21).

It is necessary to note that in other fields of occupational medicine (for example, toxicology, cancer epidemiology) studies in which the exposure has not been measured and the disease has not been assessed according to current medical standard would be given no consideration.

The diagnosis of MSDs in epidemiological studies requires a brief discussion of some peculiar aspects, which distinguish the epidemiological setting from the clinical one.

The first and foremost issue is that in the clinical setting it is the patient who seeks medical attention, usually because of symptoms which are associated to discomfort, pain or functional limitation severe enough to rise the need to bring the own situation to a doctor’s attention. Instead, in the epidemiological setting of a cross-sectional or cohort study the situation is inverted: it is the medical investigator who is looking for diseased persons in what is an otherwise (assumed) healthy working population.

This situation affects indubitably the whole diagnostic process in different ways:

symptoms are not spontaneously presented to a doctor but “prospective” patients are specifically questioned to elicit the (eventual) presence of definite symptoms: in such setting it is well known the possible presence of serious biases due, for example, to the tendency of the subject to please the investigator (reporting what was requested), or various form of recall bias (3);

the utility of physical examination (but even of instrumental tests) is drastically reduced by the low prevalence of the condition studied which affect the positive predictive value of tests with even high sensitivity or specificity (17);

expectations (even unconscious) of both the subject studied and the investigator may well impact the result of the investigation; the same may be true for the perceived utility of a “positive” study by the investigator (higher chance to publish) or by the diagnosis for the worker (compensation).

The well-known possible (and likely) presence of these serious biases has lead epidemiologists to use rigorous case definitions in observational studies and, also, to require multiple cohorts studies with a low risk of bias to accept a causal relation between an exposure and a disease (as an example, see occupational cancer epidemiology).

Based on this, one would expect that in the epidemiology of occupational musculoskeletal diseases (conditions which have a high spontaneous incidence even in non-exposed populations) case definitions were based on the best available techniques (magnetic resonance imaging, EMG, and so on), exposures were exactly measured with the appropriate instrumentation, and studies were performed with blind techniques (exposure/outcome) and other arrangements to minimize all possible biases.

Instead, an overview of the studies published in this field shows that the vast majority are performed assessing exposure and outcome by questionnaires, issue that has been already questioned (5, 21), without blinding between exposure and outcome assessment and without taking in consideration the numerous known non-occupational risk factors.

In this context, there is the need of well-planned and conducted cohort studies in which exposure and outcome are measured with the best available techniques (that is, instrumentally), with thorough consideration of all possible confounders and minimizing effectively the known possible biases: until that, even meta-analyses and systematic reviews cannot advance our knowledge in this field, as they are not substitute for well-planned and well-conducted studies.

It is therefore essential, in the analysis of the scientific literature on the relationship between certain exposures to biomechanical risk factors and MSDs, to clarify where the level of evidence of the association can be classified with reference to the scheme presented in the above table, to define, with reference for example to the pathology, if the study has investigated a disease, or a disorder, or just a symptom, without confusing the conditions in question.

The process of attributing the disease to the work-activity requires first of all to establish the level of evidence of the diagnosis which, with reference to this specific point, should be preferably equivalent to “Very probable”. A level of evidence of “Probable” (where the probability in favour of the disease may be less than 50%) may be taken into consideration when constraints (available medical technology in a given place, a decision which has to be made without the possibility to acquire other information) preclude the possibility of additional medical investigations. In this case, proper medical judgment of all the available evidence is warranted.

Once that the level of evidence of the diagnosis is defined, it will be possible to proceed to the evaluation of the exposure which should be suitable for frequency, duration and level to cause the disease, in accordance with the evidence of the scientific literature. As in the case of diagnosis, the explicit definition of the levels of evidence on the basis of which it is decided to attribute a disease to a given work activity is essential to properly frame the relevant legal consequences.

The quality of the exposure assessment, as it has been seen, has progressive levels of evidence that, without considering the situations in which the exposure can be excluded or considered unlikely, range from “possible” to “near certain”: again, with reference to this specific point, the quality of the assessment of the exposure should be preferably equivalent to “Very probable”. A level of evidence of “Probable” (where the probability in favour of a good quality of the exposure assessment may be less than 50%) may be taken into consideration when constraints (available industrial hygiene technology in a given place, a decision which has to be made without the possibility to acquire other information) preclude the possibility of additional workplace investigations. In this case, proper professional judgment of all the available evidence is warranted.

The easiest case for a positive attribution is one in which we are faced with a well-defined disease under the diagnostic profile and exposure data of the worker demonstrating that the same, by frequency (daily?), duration (the conventional 40 hours per week for a sufficient number of years?) and level (above the acceptable limits?), possess the necessary and sufficient power to cause the disease in question.

In other cases, however, in the presence of a well-defined disease, we can find ourselves faced with a lack of documentation of the exposure, for example because the company where the worker has operated no longer exists and/or no assessments are available. Given that the quality of information about the exposure, in this case, is less than optimal, it will be possible to refer to epidemiological criteria (evidence that those who carry out tasks such as those of the worker in question - or have worked in companies of that type - have a significantly increased risk of or other evidence available in the concrete case). In this situation, however, the evidence of exposure can never reach the level of “Very Probable”.

Weighing of non-occupational risk factors in attribution

For the attribution of a disease to a working exposure the evaluation of non-occupational risk factors to the disease is also required. Non-occupational risk factors, such as age, body-mass-index, co-morbidities (e.g., diabetes) or non-occupational activities may alone be a cause of MSDs. The quality of the epidemiologic evidence linking these factors to a specific MSD should be considered with the same rigor as the evaluation of the literature on occupational factors. In addition, these factors may be present in an individual on a spectrum, i.e., almost as a dose, and that spectrum should be considered in the interpretation of the literature and the assessment of the individual.

It is important to note that the presence of a non-occupational risk factor (such as age, sex, obesity, co-morbidities, and so on) does not rule out, per se, causality for an occupational exposure. The role of occupational exposures in an individual worker must be considered to decide if occupational exposure substantially increase the risk of an MSD in the individual worker. The presence of non-occupational risk factors may be roughly classified as:

“dominant” (in the specific individual they are present in a relevant measure, such that the individual may be classified as “high” risk, with respect to the average distribution of such risk factors in the general population);

“average” (in the specific individual they are present in a measure similar to the average distribution of such risk factors in the general population);

“non dominant” (in the specific individual they are present in a reduced measure, such that the individual may be classified as “low” risk, with respect to the average distribution of such risk factors in the general population).

Minimum latency from start of exposure

This is what is called “Induction Time”, i.e. the time elapsed from the onset of exposure to the time the disease began. The minimum latency, if known, may be a useful criterion of attribution exclusion. Obviously, the date of onset of the disease may have been well before the diagnosis was made by a physician. Minimum latency is obviously only a consideration for diseases based on a model of cumulative exposure and not for an acute injury. For example, the well-established natural history of carpal tunnel syndrome in pregnancy (24) shows that this disease usually appears after the fourth month of pregnancy: for analogy, it may well be assumed that a period of only a few months of working activity which requires extremely forceful, fast and continuous hand exertions may be the cause of carpal tunnel syndrome in a female worker. On the contrary, as osteoarthritis requires years to develop: if this disease is diagnosed in a worker only a few months after the start of a working activity, the attribution of the disease to that activity is easily ruled out.

Maximum latency from end of exposure

If known, it is a useful criterion for the exclusion of attribution: exposures too far from the date of onset of the disease (for example, years before the diagnosis of carpal tunnel syndrome) are probably not to be considered causally relevant. However, good data on this issue for occupational MSDs are lacking.

Overall assessment

The overall assessment of probability of an occupational MSD needs to integrate the likelihood estimate of the disease with the estimated attributable contribution of the workplace exposure and the contribution of non-occupational factors. An example of how these might be combined to form an overall probability estimate is presented in the following table. Table 4 is just an example: the probability shown in each cell will vary depending on the interpretation of the available data. In addition, the table does not consider all the different possible combinations of factors. Moreover, it should be noted that as the likelihood of a substantial occupational exposure increase, the role of non-occupational risk factors in the attribution of the disease to the occupational activity may be considered less relevant (up to indifferent).

Table 4

Examples of the level of evidence of attribution of the disease to the occupation, based on the level of evidence of the disease, occupational exposure and non-occupational factors

Disease	Occupational exposure (causally relevant)	Non-occupational factors (causally relevant)	Attribution of the disease to the occupation
Insufficient orNot likely	Insufficient orNot likely	Dominant	Ruled out orNot likely
Possible	Possible	Dominant	Not Likely
Probable	Probable	Average	To be evaluated on the basis of the specific evidence available
Very probable	Very probable	Non dominant	Very probable
Near certain	Near certain	Non dominant	Near certain

Discussion

After decades of research, the field of musculoskeletal diseases in occupational medicine still presents critical gap in information.

Whereas there are a number of studies where both exposure and outcome are defined by means of questionnaires, we have almost no study where outcome has been defined objectively (with some exceptions for carpal tunnel syndrome and lumbar disc disease). Also, there are almost no study where exposure has been measured instrumentally and, as a consequence, the dose-effect relationship between biomechanical risk factors and MSDs remains largely unknown.

This makes the attribution of an MSD to occupation essentially subjective, as demonstrated by the large variations in national statistics among similar countries, which cannot be attributed to real differences in exposure prevalence and intensity.

The criteria proposed in this paper may help the occupational physician in achieving a greater consistency in attributing occupational diseases to biomechanical risk factors: in addition, they make the decisional process less opaque, offering the possibility to attribute a specific level of quality to the assessment of the disease and of the exposure and they also offer an explicit framework for attributing a specific disease to a workplace exposure, also taking into account non-occupational factors which may be causally relevant.

However, to advance the current knowledge about biomechanical risk factors (to the level, for example, reached for a number of physical and chemical agents) we need high-quality, cohort studies in which exposure is measured instrumentally and outcome is assessed objectively, with the best available appropriate techniques (EMG, MRI, other), and complete control of confounders and possible sources of bias.

These studies are the only ones which can provide accurate (enough) information to establish a reliable (enough) dose-effect relationship, which could form the basis for the causal attribution of musculoskeletal diseases and for the proposal of occupational exposure limits to prevent musculoskeletal diseases (and disorders, and, if possible, even symptoms).

It is the responsibility of researchers to perform studies which effectively advance the knowledge available at a certain time, however difficult it may seem; it is the responsibility of funding agencies to make possible to researchers to undertake these studies and it is the responsibility of scientific journals’ editors to select for publication only studies which produce the knowledge in need.

No potential conflict of interest relevant to this article was reported by the author

Tabella 1

Diversi criteri per la qualità (e la probabilità associata) della diagnosi di tendinopatia della spalla

Criteri di diagnosi	Livello di evidenza
Dolore in regione scapolare al collo con dolorabilità alla palpazione del muscolo trapezio.	Insufficiente(può essere solo disagio transitorio)
Dolore alla spalla peggiorato dall’abduzione del braccio.	Possibile
Dolore alla spalla con rilievo di dolorabilità alla palpazione dell’acromion e dolore con abduzione volontaria del braccio superiore a 45 gradi (13).	Probabile
Sintomi, rilievi clinici e risonanza magnetica che riporta ben definite alterazioni del segnale tendineo nel lato interessato e (se è il caso) assenza di alterazioni nella spalla asintomatica.	Molto probabile
Relazione di intervento chirurgico con documentazione di degenerazione o lacerazione tendineo e istologia che dimostrala tendinosi.	Praticamente certa

Tabella 2

Criteri per la qualità (e la probabilità associata) dell’esposizione a fattori di rischio biomeccanici (indipendentemente dall’intensità)

Valutazione dell’esposizione	Livello di evidenza
Esposizione descritta dal lavoratore o presunta in base a una mansione lavorativa.	Insufficiente (l’esposizione ai rilevanti fattori di rischio biomeccanici non può essere valutata oggettivamente sulla base solo di una descrizione soggettiva o di una mansione)
Documentazione amministrativa dell’impiego presso un’azienda, mansione e qualche documentazione scritta sul lavoro da cui è possibile assumere l’esposizione, almeno qualitativamente.	Possibile
Valutazione scritta da parte di un valutatore esperto che ha osservato il lavoro e documentato le esposizioni biomeccaniche mediante una checklist appropriata o uno strumento simile.	Probabile
Analisi video con alcune misurazioni dei fattori di rischio biomeccanico (ad esempio, peso dello strumento o delle parti, stima delle posture, cicli di attività e così via).	Molto probabile
Misurazione dei movimenti del corpo o di parti del corpo del lavoratore mediante registrazioni video dettagliate e/o unità di misura inerziali, misurazioni della forza esercitata per mezzo di sensori meccanici durante l’esecuzione del compito e così via.	Praticamente certa

Tabella 3

Schema per classificare la qualità degli studi che indaghino la relazione tra l’esposizione a fattori di rischio biomeccanici e le malattie muscolo-scheletriche

			Valutazione dell’esposizione
			Valutazione obiettiva		Valutazione indiretta
			Metodi quantitativi di misurazione diretta	Analisi video o osservazioni basate su video o protocolli sperimentali	Osservazioni di esperti	Mansione, valutazione auto-riferita, matrice mansione-esposizione
Definizione di caso	Criteri diagnostici	Imaging (più esame fisico)	++/++	++/+	++/-	++/--
		Esame fisico (sintomi più segni clinici)	+/++	+/+	+/-	+/--
	Sintomi	Intervista strutturata (storia clinica attuale e passata)	-/++	-/+	-/-	-/--
		Questionario autosomministrato	--/++	--/+	--/-	--/--

Tabella 4

Esempi del livello di evidenza dell’attribuzione della malattia all’occupazione, in base al livello di evidenza della malattia, esposizione occupazionale e fattori non occupazionali

Malattia	Esposizione occupazionale(causalmente rilevante)	Fattori di rischio non occupazionali(causalmente rilevanti)	Attribuzione della malattia all’occupazione
Insufficiente o Improbabile	Insufficiente o Improbabile	Dominanti	Esclusa o Improbabile
Possibile	Possibile	Dominanti	Improbabile
Probabile	Probabile	Nella media	Da valutare sulla base della specifica evidenza disponibile
Molto probabile	Molto probabile	Non dominanti	Molto probabile
Praticamente certa	Praticamente certa	Non dominanti	Praticamente certa