UNIVERSITAT AUTÓNOMA DE BARCELONA FACULTAT DE … · in Figure 2. Results are presented in two...

13
UNIVERSITAT AUTÓNOMA DE BARCELONA FACULTAT DE MEDICINA DEPARTAMENT DE FARMACOLOGIA I PSIQUIATRÍA DIVISIÓ DE FARMACOLOGÍA APROXIMACIÓ A L'ESTUDI DE LA PROBLEMÁTICA ACTUAL DEL TRACTAMENT DEL DOLOR Tesi presentada per Fèlix Bosch i Llonch per a optar al grau de Doctor en Medicina i Cirurgia, i dirigida pel Dr. Josep-Eladi Baños i Diez. Bellaterra, setembre del 1992

Transcript of UNIVERSITAT AUTÓNOMA DE BARCELONA FACULTAT DE … · in Figure 2. Results are presented in two...

UNIVERSITAT AUTÓNOMA DE BARCELONA

FACULTAT DE MEDICINA

DEPARTAMENT DE FARMACOLOGIA I PSIQUIATRÍA

DIVISIÓ DE FARMACOLOGÍA

APROXIMACIÓ A L'ESTUDI DE LA

PROBLEMÁTICA ACTUAL DEL

TRACTAMENT DEL DOLOR

Tesi presentada per Fèlix Bosch i Llonch pera optar al grau de Doctor en Medicina iCirurgia, i dirigida pel Dr. Josep-EladiBaños i Diez.

Bellaterra, setembre del 1992

Meih and Find E\p Clin Pharmacol 1989; 11(2): 123-127

Acceptability of Visual Analogue Scales in the Clinical Setting: AComparison with Verbal Rating Scales in Postoperative PainJ.E. Baños1, F. Bosch1, M. Cañellas2, A. Bassols2, F. Ortega2 and J. Bigorra*

'Department of Farmacologia i Psiquiatria, Universitat Autònoma de Barcelona; 2Servei d'Anestesiologia,Reanimació i Clínica del Dolor, Hospital de Sabadell; 3Servei de Farmacologia Clínica, Hospital Clínic i Provin-cial, Barcelona, Spain

SUMMARYPain is the clinical symptom most difflculi to evaluate. Although clinical trials methodology have permitted assessment of pain objectively

through rating scales, these strategies hu\e not been used in clinical selling. The present study was undertaken to determine if visual analoguescales could be useful in the measurement of postoperative pain in usual medical practice. The study comprised 212 patients with abdominal,orthopedic or gynecological surgical procedures within the previous 24 h. Patients evaluated their pain using a verbal rating scale (VRS) of fivepoints or a visual analogue scale (VASi of 10 cm. The investigators also evaluated patient pain through a VAS. The results obtained showedthat a high correlation between VRS and VAS could be established in a/I patients (p<OMl). The VAS of patients and researchers were alsofound to be highly correlated <p<0.00li. When values of each group were compared by pain intensity a total agreement of VAS scores at lowpain level could be established, but differences Kere found at high pain intensity levels, suggesting that physicians scored lower than patientswhen pain was severe to unbearable. It is concluded that VAS could be a reliable method to assess pain in clinical setting.

Key words: Postoperative pain - Pain scales - Visual analogue scale

INTRODUCTIONIt is well known that pain is one of the clinical symp-

toms more difficult to evaluate. Its subjective characterand the complex feelings that pain evokes make dif-ficult its accurate measurement by physicians andnurses, and, as suggested by Huskisson (1), perhaps on-ly the patients can quantify it correctly. However,several clinical trials on analgesics have shown that painscales can be a reliable and objective method to assessthe analgesic properties of drugs. It is generally acceptedthat visual analogue scales are more sensitive and ac-curate than other measures, thus being generally ad-vised in the evaluation of the intensity of pain.

The problem arises when quantification of pain mustbe done in the clinical setting. It has been suggestedthat sensitivity to pain is higher under clinical thanunder experimental conditions (2), and it is possible thatclinical trial conditions could also alter the efficacy ofanalgesic therapy when compared with commonmedical practice. Additionally, neither practicing physi-cians nor nurses are generally aware of the use of painscales. Their therapeutic decisions rely on subjective

evaluation of pain after asking the patient about it, andanalgesics are often prescribed without considering theactual intensity of pain in every case. Moreover, theconditions of clinical trials are frequently not equivalentto those in the usual clinical setting, and methodologiesused in the former could be useless when applied to thelatter. For these reasons, there is an urgent need to testwhether pain scales can be as advantageous in every-day medical practice as they are in clinical trials. Somepreliminary reports suggest that visual analogue scalescould be useful in assessing pain in these conditions (3).

The present study was undertaken in order toestablish the usefulness of a visual analogue scale inthe measurement of the intensity of postoperative painin usual medical practice.

PATIENTS AND METHODS

Patient selectionTwo hundred and twelve patients with abdominal,

orthopedic or gynecological surgical procedures withinthe previous 24 h were admitted to the study. They were

0379-0355/89/$5.00 + $2.00Cop>ii i ih i 1989 J.R. Prous, S.A.

124 J.E. Baños et al.

recruited from three hospitals (A, B and C), two ofthem (A and B) with General Surgery (GS) Departmentsand the third (C) with Departments of Obstetrics andGynecology (OG) and Traumatology and Ortophedics(TO). Only patients older than 10 years and able tofulfil l pain scales were considered eligible. Since thisstudy was intended as a naturalistic evaluation, no otherinclusion and exclusion criteria were used. The mainfeatures of the study population are shown in Table 1.

TABLE 1. Main characteristics of Ihc patients

Percentage

HospitalABC

Age (years)<3031-50>50

SexMF

Surgical ProcedureGSTOGO

4838

126

4369

100

105107

8610323

22.617.959.4

20.332.547.2

49.550.5

40.648.610.8

Pain assessmentPain was assessed both by the patient and the physi-

cian always within the first 24 h after the surgical pro-cedure. Patients were instructed by a trained interviewerto rate the intensity of pain, if any, by using a five itemverbal rating scale (VRS, 0 = no pain; l = mild pain;2 = severe pain; 3 = very severe pain; 4 = unbearablepain) and a 10 cm horizontal visual analogue scale(VAS) with the statements «no pain» and «unbearablepain» at its extremes. The VAS was always administeredimmediately after the VRS. At the same visit and beforethe patient assessments, the responsible physicians (MC,AB and FO) evaluated the severity of pain by using anidentical VAS. The physician was always blind to thepatient scores, as well as to the administered analgesicmedication before evaluating pain.

Anesthesia and postoperative analgesiaInformation on the anesthesia and drugs used dur-

ing the procedure (either inhaled, intravenous or local)was recorded. These data are not presented here since

the aim of the study was focused on the severity of painregardless of the analgesic received. The investigatorsdid not influence the medication prescribed for reliefof pain and the staff responsible for prescription ofanalgesic drugs was intentionalh kept unaware of thisstudy being conducted in order not to alter the usualmedical practice.

Data analysisThe overall correlation between VRS and VAS

recorded by the patient was evaluated using the Spear-man's ranked correlation test. The correlation betweenVAS scores according to patient and investigator'sassessment was established by using the Pearson's cor-relation test. A separate evaluation by center andmedical speciality was also performed. A value ofp<0.05 was considered statistical!) significant.

RESULTS

Pain severity valuesThe results obtained by using the VRS in the three

hospitals are summarized in Figure 1. A wide range ofpain intensity scores was observed. Important dif-ferences among hospitals were seen, especially withhospital A, where a high number of patients (89.6%)reported mild or no pain. When this finding was in-vestigated, we found that the majority of surgical pro-cedures performed over the study period were knownas basically painless procedures. The variability observ-ed in hospitals B and C was within the expected range.

Overall, pain was absent in 15.5% of the patients,mild in 58.6%, severe in 19%, \ery severe in 4% andunbearable in 2.9%. Taking into account that this wasa naturalistic study, it is important to consider the largenumber of patients with high severity scores sincealmost 25% of the patients described their pain as severeto unbearable.

The VAS values recorded by the patients are depictedin Figure 2. Results are presented in two point inter-vals except for the lowest range, where the values from0 to 0.2 cm were considered equivalent to 0 cm, thusbeing the second interval 0.3-2 cm.

VAS versus VRS in patientsThe correlation between different values of the scales

and some characteristics of the patients is shown inTable 2. When patients were considered globally, a highcorrelation (p<0.001) was observed. Sex, age, type ofsurgical procedure and hospital center did not influencethe correlation between both scales. The investigators

Pain sca/es in the clinical setting 125

Percentage

48-

38-

28

18-

~

1

J~VRS SCORE S|•

j'

*" ¡-. . ,

Mm CENTER A

D CENTER B

Q CENTER C

H i ~[H lïfl EI

FIG. I . VRS scores by center. 1 = no pain, 2 = mild pain, 3 = severe pain, 4 = very severe pain, 5 = unbearable pain. Vertical bars represent percen-tage of patients with each score in each hospital center.

had no difficulties in obtaining the VAS scores. WhenVRS values were compared with results obtained byVAS, no pain (15.5%) appeared to be equivalent to the0-0.2 interval (16.4%), mild pain (58.6%) to the 0.3-4interval (61%), severe pain (19%) to the 4.1-6 inter-val (13.2%), very severe pain (4%) to the 6.1-8 inter-val (5.1%) and unbearable pain (2.9%) to the highestVAS values 8-10(3.1%).

VAS values of patients versus VAS values ofinvestigators

The correlation between VAS values scored by thepatients and the investigators is summarized in Table

3. Again, no differences were observed due to age, sex,surgical procedure or hospital center, although the cor-relation was poorer in patients undergoing obstetric andgynecologic procedures (r = 0.75). This finding couldbe due to the small sample size. An analysis by hospitalcenter showed that hospital B had the lowest correla-tion coefficient (r = 0.66). When VAS values scored bypatients were compared with those scored by the in-vestigators, a noticeable trend was found (Table 4). Atlower VAS values, there was almost a coincidence be-tween both values, but as pain intensity increased, thedifferences became progressively larger, up to 30% athigher VAS values.

Percentage

of

rat lent s

e z 2-4 i s 6-8 e-ie

FIG. 2. VAS scores by center. Vertical bars represent percentage of patients with scores in each interval in each hospital center.

126 J.E. Baños et al.

TABLE 2. Correlation bel» ten VAS and VRS

SexMaleFemale

Age (y)<3031-50>50

Surgical ProcedureGSTOGO

HospitalABC

•In all cases, p<0.001.

TABLE 3. Correlation

SexMaleFemale

Age (y)£3031-50>50

Surgical ProcedureGSTOGO

HospitalABC

N

10510-

4369

100

8610323

4838

126

Spearman's ranked correlation

between \ AS \alues scored b>investigators

N

105107

4369

100

8610323

4838

126

Correlationcoefficient"

0.730.83

0.660.750.85

0.870.710.69

0.890.80o-o

lest.

patients and

Correlationcoefficient"

0.830.92

0.780.880.88

0.870.890.75

0.960.660.88

TABLK 4. VAS values scored h} patients and imestigalors in the stud>centers

VAS scored ^S scored by the imestigators"

b\ the patients Centre A Centre B Centre C

•In all cases, p<0.001, Pearson's correlation test.

DISCUSSIONIn recent years there has been an increasing interest

in the accurate evaluation of pain. As Wallenstein (4)pointed out, casual or random measurements by nursesor physicians are no longer acceptable. Clinical trialson analgesics have provided the necessary framework

123456789

10

0.7 (70)1.6 (80)2.5 (76)3.4 (85)4.2 (84)5.0 (83)5.9 (84)

———

2.0 (NE)2.4 (NE)2.9 (97)3.4 (85)3.9 (76)4.4 (73)4.9 (70)5.4 (68)

-

1.0 (100)1.8 (90)2.5 (83)3.2 (80)3.8 (76)4.5 (75)5.2 (74)5.9 (-4)6.8 (75)7-3 (73)

•Figures in brackets refer to the percentage of agreement (PA) w i t hpatients' values calculated as PA = (VAS investigator/VAS pa-tient).\100. NE = not evaluated.

and methodology to scientifically evaluate pain (5, 6).However, this fact has not been followed by an im-provement of pain therapy in usual clinical practicesince a large number of patients are still suffering frompain in spite of analgesic treatment (7-9). Althoughsome wrong assumptions in pain therapy and a limitedknowledge of analgesic pharmacology could explain theinadequacy of pain relief, there is no doubt that a cor-rect evaluation might enhance the probability of an im-proved treatment.

Among the potential methods useful in pain assess-ment, visual analogue scales (VAS) are commonly con-sidered more sensitive and accurate than other ratingscales (1, 10-12), as they allow a continuous evaluationwhich is not feasible by using verbal rating scales (VRS).In fact, VRS only-use ranked values which are notnecessarily equivalent to an identical difference in painseverity. Moreover, the verbal descriptors commonlyused in VRS could have different meanings for differentpeople (6). The present study compared a standard VRSwi th a well known VAS. We used the VAS proposedby Huskisson (1) and Wallenstein (4). The results ob-tained were similar to those reported in previous clinicaltrials (5, 13). The observed agreement is an interestingfinding as it suggests that VAS could be a useful toolin the evaluation of pain in everyday practice. VAS isan accurate, sensible, cost-effective and easy-to-usemethod that allows the measurement of a subjectivesymptom. This is a very important consideration whenprescribing analgesic therapy, as pain severi ty is themost important factor on which the selection of ananalgesic drug should be based. Other studies are now

Pain scales in the clinical setting 127

in progress in the same hospitals to evaluate if wideruse of VAS can result in an improved pain relief inpostoperative patients.

A question frequently neglected in analgesic studiesis the values scored by the investigator and patient. Ourresults suggest that there is a good correlation betweeninvestigator and patients' assessments. However, as painseverity increases, researchers are prone to score lowerthan patients. This is in agreement wi th the results ob-tained by Hodgkins el at. (14) in rheumatic patients,although they reported such a trend in all pain severities.However, their differences were not statistically dif-ferent. We did not perform any statistical analysis ofthis subset of patients since the sample size was toosmall. If confirmed, this finding would suggest thatphysicians undertreat pain due, at least to some extent,to their perception that patients have less pain that theyclaim. If this is so, measures should be taken to avoidunnecessary discomfort for patients in pain.

REFERENCES1. Huskisson, E.G. Measurement of pain. Lancet 1974; 2:

1127-1131.2. Anonymous. Sensitivity to pain greater in a clinical than in a

laboratory setting. JAMA 1983; 718.3. Joyce, C.R.B., Zutski, D.W., Hrubes, V and Mason, R.M.

Comparison affixed interval and visual analogue scales for ratingchronic pain. Eur J Clin Pharmacol 1975; 8: 415-420.

4. Wallenstein, S.L. The evaluation of analgesics in man. In:Analgesics: Neurochemical, Behavioural and Clinical Perspec-

tives. M. Kuhar and G. Pasternak (Eds.). Raven Press: New York1984; 235-255.

5. Wallenstein, S.L., Heydrich III, G., Kaiko, R. and Houde, R.W.Clinical evaluation of mild analgesics: The measurement ofclinical pain. Br J Clin Pharmacol 1980; 10: 319S-327S.

6. Bigorra, J. Evaluación de fármacos analgésicos en el hombre.Dolor 1988; 3: 40-44.

7. Marks. R.M. and Sachar, E.F. Undcrtrealment of medical in-patients with narcotic analgesics. Ann Intern Med 1973; 78:173-181.

8. Cohen, F. Postsurgicalpain relief: Patients, status and nurses'medication choices. Pain 1980; 9: 265-274.

9. Donovan, M., Dillon, P. and McGuire, L. Incidence andcharacteristics of pain in a sample of medical-surgical inpaiients.Pain 1987; 30: 69-78.

10. Scott, J. and Huskisson, E.C. Graphic representation of pain.Pain 1976; 2: 175-184.

11. Sriwanatakul, K., Kehie, W. and Lasagna, L. The quantifica-tion of pain: An analysis of words used to describe pain andanalgesia in clinical trials. Clin Pharmacol Ther 1982; 32: 132-148.

12. Seymour, R.A. The use of pain scales in assessing the efficacyof analgesics in post-operanve dental pain. Eur J Clin Fharmacol1982; 23: 441-444.

13. Ohnhaus, L.E. and Adler, R. Methodological problems in ¡hemeasurement of pain: A comparison between the verbal ratingscale and the visual analogue scale. Pain 1975; 1: 379-384.

14. Hodgkins, M., Albert, D. and Daltroy, L. Comparing patients'and their physicians'assessments of pain. Pain 1985; 23: 273-277.

Address all correspondence to: J.E. Baños, Departament de Far-macologia i Psiquiatria, Facultat de Medicina, Universitat Autònomade Barcelona. 08193-Bellaterra (Barcelona), Spain.

The Clinical Journal of Pain6:206-211 © 1990 Raven Press, Ltd., New York

A Survey of Pain Complaints and Treatment by GeneralPractitioners in the Spanish Public Health Organization

Fèlix Bosch, M.D., *Ignacio Toranzo, M.D., and Josep E. Baños, M.D., Ph.D.

Departament de Farmacologia i Psiquiatria, Facultat de Medicina, Universitat Autònoma de Barcelona, and *Grupd'Estudis de la Salut, Sabadell, Spain

Abstract: The therapeutic habits of general practitioners are an important cluewhen drug therapy is considered, because they are treating the most frequentcomplaints. When pain problems are considered, it would be valuable to de-termine the characteristics of the pain consultations and their therapeutic at-tempts to solve these complaints. The present study was designed to elucidatethe characteristics of pain diagnoses and treatment approaches at primary-carelevel in Spain. A total of 299 patients were evaluated, considering pain loca-tion, diagnostic syndrome, previous therapies, and treatments selected by the13 participating physicians. Limb and back pain were the most frequent paincomplaints. A third of the patients had received previous treatment and 36%were self-medicating, mainly with aspirin or paracetamol. Physicians pre-scribed diclofenac at full doses, but aspirin and paracetamol were used atsubtherapeutic dosages. The study showed that (a) rheumatic pain was themost frequent at primary-care level, (b) a high level of self-medication wasdetermined, therefore recommending a careful drug history, and (c) miscon-ceptions about analgesic drugs may partially explain the therapeutic failure insome patients. Educational programs in rheumatic pain and analgesic therapyfor general practitioners are strongly recommended. Key Words: Epidemiol-ogy—Rheumatic pain—Pain treatment—Primary care—Self-medication—Analgesics.

In most countries, general practitioners providethe first level of medical care. Pain is one of themore frequent symptoms that patients complainabout to their physicians (1). However, the charac-teristics and treatment of such complaints in generalpractice has been a relatively neglected issue. Onlyvery recently have epidemiologic studies on paincomplaints in the general population published (1-4). Moreover, the core of work on pain problems

Address correspondence and reprint requests to Dr. J. E.Baños, at Departament de Farmacologia i Psiquiatria, Facultatde Medicina, Universitat Autònoma de Barcelona, 08193 Bel-laterra, Barcelona, Spain.

has been carried out by medical specialists (anes-thesiologists, rheumatologists, neurologists) and bythose working in pain clinics. This fact, though log-ical, has sometimes resulted in a biased approach topain problems (I). Therefore, although considerableeffort has been expended to facilitate pain treatmentin the past, little can be achieved without a substan-tial change in the attitudes and prescription habitsof general practitioners, because most human suf-fering is actually in their hands. They visit nearly allpatients, treating them for prolonged periods, andare requested to manage a great number of paintypes. It would therefore be instructive to designstudies on pain and its management in primary

206

PAIN A T PRIMAR Y CARE IN SPAIN 207

health care, because they could offer an extensiveand realistic picture of the present stage of painproblems in a community.

The present study was undertaken to establishhow general practitioners are coping with pain inSpain. The purposes of this survey were (a) to studythe characteristics of pain in patients who demandmedical care for this symptom, (b) to evaluate howthis pain is treated before asking for medical care,and (c) to define the therapeutic measures selectedby primary care physicians.

METHODS

Data were collected by 13 general practitionersworking in the Primary Health Centers (PHC) ofInstitut Català de la Salut, the public health organi-zation of Catalonia. These PHCs were located inSabadell, a city of nearly 200,000 inhabitants. Pa-tients were recruited when they attended PHC orrequested domiciliary assistance for medical care ofpain complaints. Patients were never seen in a pri-vate practice setting.

Three hundred eleven patients were recruited. Acase form was completed at the time of the first visitfor each of them. Initial data included age and sex,type of assistance (domiciliary or PHC), evaluationof pain (diagnosis, duration, and intensity), andtreatments (previous treatment, self-treatment ifany, and therapeutic measures selected by the phy-sician participating in the study).

Evaluation of painOnly diagnostic syndromes were established,

(e.g., headache, but not cluster headache, tensionheadache, or migraine), because an accurate diag-nosis is not usually achieved at the primary-carelevel. Thereafter, each type of pain was classifiedon the basis of its location. Pain was considered"acute" when its duration was less than 1 week and"persistent" if longer.

Evaluation of intensity of painThe evaluation of pain intensity was performed

using a visual analogue scale (VAS) of 10 cm with"no pain" and "unbearable pain" at its ends. Bothpatients and physicians independently completed aVAS. Physicians scored first on one side of the caseform and then the patient on the other, both of thembeing unaware of the score awarded by the other. A

verbal rating scale (VRS) of five categories (slight,moderate, severe, very severe, excruciating) wasalso completed by the patients before VAS scoring.

Statistical analysisData were analyzed by means of Statistical Pack-

age for Social Scientists (SPSS) software. Correla-tions between VAS scores of patient and physician,and between VAS and VRS of patients, were eval-uated by Spearman and Pearson correlation tests,respectively.

RESULTS

Characteristics of pain complaints

General characteristicsTwelve (4%) of the 311 patients accepted in the

study were excluded either for lack of cooperationor for their inability to understand the use of theVAS. For these reasons, only 299 patients wereconsidered in the final analysis. The mean age was43.5 years (SD ± 18.4, range 14-90) and 61% werewomen. A majority of them attended the consultingroom at PHC. Table 1 shows a more detailed de-scription of the sample.

Distribution of pain complaintsTable 2 shows the frequency of the different pain

complaints classified by pain location and sex. Themost frequent diagnoses were limb and back pain,reported by 31.1% and 25.4% of the patients, re-spectively, followed by head, face, and mouth pain(15.1%) and abdominal pain (12%). The most fre-quent diagnostic syndrome was back pain (31.1%)followed by peripheral joint pain (14%). When eval-uated by sex, headache and peripheral joint painwere more frequent in women, whereas musculo-skeletal and traumatic pain were more common inmen.

The diagnoses designed as "other" in Table 2 and

TABLE 1. Characteristics of the sample

Characteristics

M/FAge group (yr)

<2425̂ 1445-64>65

Type of assistancePrimary health centerAt home

N

117/182

5110010543

25841

%

39.1/60.9

17.133.435.114.4

86.313.7

The Clinical Journal of Pain. Vol. 6. No. 3. 1990

208 F. BOSCH ET AL.

TABLE 2. Prevalence of pain location by sex

Men

LimbsBackHead. face, mouthAbdominalChestOther*

" Percentage by sex6 See text.

N

4330121589

%"

36.725.610.312.86.87.7

WomenN

504633219

23

%"

27.525.318.111.54.9

12.6

TotalN

937645361732

%

31.125.415.112.05.7

10.7

following comprised 32 patients who had psycho-logical (10 patients), throat (seven patients), cancer(six patients), and ischemic limb pain (five pa-tients), and a group of miscellaneous diagnoses(four patients).

Intensity of painTable 3 provides the intensity of pain in relation

to the most frequent location as measured by a VRSscale. In our study, pain was experienced as mostintense when related to head, face, and mouth,whereas the mildest was chest pain. The most pain-ful diagnostic syndromes were headache and odon-talgia: more than half of the patients reported theirpain as very severe or excruciating. Colic pain,classified in the abdominal group, was also verypainful.

When considering the type of assistance (at homeor PHC). the majority of patients (90%) assisted athome rated their pain from severe to excruciating.On the contrary, in ambulatory patients the inten-sity was clearly lower, even though 80 of these pa-tients (26.7%) scored their pain from very severe toexcruciating. A high correlation was observed whenindividual VRS and VAS scores were observed.Good correlation was also obtained when VASscores of physicians and patients were calculated;although at high intensity values physicians tendedto rate the pain lower than the patients.

Duration of painWhen considering duration, 60% of the patients

defined their pain as acute, whereas the remainingdescribed it as persistent. Acute pain was more fre-quent in nearly all the diagnostic syndromes, butpersistent pain predominated in several groups ascancer, psychological pain, back, and peripheraljoint pain. In the group of headache, the number ofpatients developing acute of persistent pain wasnearly equal.

Characteristics of pain treatment

Previous medical treatment of painA significant number of patients had received

previous medical attention for their pain complaint.A total of 113 treatments (drugs and other therapeu-tic measures) were prescribed to 85 patients, result-ing in an average 1.3 treatments per patient. Table 4summarizes these previous treatments. As shown,the most frequently prescribed drugs were nonste-roidal anti-inflammatory drugs (NSAID, 36.3%) fol-lowed by paracetamol (9.7%) and topical drugs(8.8%). A similar administration pattern was ob-served when treatments prescribed by the physi-cians participating in the study were evaluated (Ta-ble 4). In this group, only a trend to prescribe moretopical drugs and nondrug therapies was remark-able, but the average treatments received by eachpatient were similar (1.4) to the value observed inthose treated previously (see above).

Self-medicationNearly a third (107) of all the patients reported

the intake of drugs by self-prescription. A total of126 drugs, that is, 1.2 per patient were used. In thiscategory, neither sex (37% men, 63% women) norage (43 ± 17.6 years) distribution were differentfrom those obtained in the general group of thestudy. As shown in Table 5, most of the self-

TABLE 3. Intensity of pain evaluated by the verbal rating scale

LimbsBackHead, face, mouthAbdominalChestOther0

Total

Slight

9(9.7)2 (2.6)1 (2.2)2 (5.6)1 (5.9)2 (6.3)

17 (5.7)

Moderate

27 (29.0121 (27.6)15 (33.3)13(36.1)8(47.1)

10(31.2)94(31.4)

Severe

34 (36.6)26 (34.2)10 (22.2)9 (25.0)4 (23.5)9(28.1)

92 (30.8)

Very severe

17(17.2)20 (26.3)1 1 (24.4)9 (25.0)3 (17.6)9(28.1)

69(23.1)

Excruciating

6(7.5)7 ( 9 . 2 )8 (17.8)3(8.3)1 (5.9)2 (6.3)

27 (9.0)

Numbers in parentheses represent percentage of verbal rating scale scores by type of pain.° See text.

The Clinical Journal of Pain. Vol. 6. No. 3. 1990

PAIN AT PRIMAR Y CARE IN SPAIN 209

TABLE 4. Types of therapies in previous prescriptionand physician selected therapies

TABLE 6. Types of self-administered treatments (%)"

Acetic acid derivatives"Topical compoundsParacetamolPropionic acid derivatives*SalicylatesVitamin mixtures'SpasmolyticsPiroxicamDigestive drugs'*MetamizolCodeine mixturesVascular drugs'Ergotic drugsOpioid drug/Other drugsOther treatments

Previousprescription

N = 113

16.88.89.75.37.16.22.64.42.62.64.41.80.92.6

21.21.8

Physician-selectedtherapiesN = 421

14.014.511.68.66.94.33.63.13.12.12.11.71.41.2

14.76.9

Results are expressed in percentages.° Diclofenac, indomethacin, and fentiazac.* Naproxen, ibuprofen, and ketoprofen.r With and without analgesics.d Includes antacid, antiemetic, and antiulcer drugs.' Includes antihemorrhoid and antivaricose agents.1 Buprenorphine and dextropropoxy phene.

medicated patients had head, face, and mouth pain,the incidence being higher in those with otalgia(58.3%), odontalgia (57.1%), and headache (47.3%).Other pain groups, such as pain of psychologicalorigin, also showed a high frequency of self-medication (near 60%). Abdominal and chest painwere the syndromes inducing self-medication to alesser extent (<20%). Table 6 shows the drugs mostfrequently used. Over-the-counter analgesics, sali-cylates, and paracetamol comprised nearly 54% ofall self-medication drugs, but it is interesting to notethe high use of metamizol (7.1%) and other NSAIDs(8%), mainly diclofenac. Drugs were usually admin-istered by the oral route, but also by topical and

TABLE 5. Self-medication by types of pain

Head, face, mouthLimbsBackChestAbdominalOther*Total

N

243327

36

14107

53.335.735.517.616.743.435.8

SalicylatesParacetamolTopical compoundsMetamizolAcetic acid derivativesErgotic drugsVascular drugsDigestive drugsPropionic acid derivativesPiroxicamCodeine mixturesSpasmolyticsOther drugsOther treatments

31.722.213.57.14.82.42.41.61.61.61.60.84.82.4

" Percentage of all patients in each category.* See text.

° Drugs considered in each group are described in the foot-notes of Table 4.

rectal administration. Intramuscular injection wasnot used at all.

Treatments prescribed by the physicians ofthe study

Table 7 shows the distribution of treatments de-pending on pain location. It is interesting to pointout the high use of NSAID in limb and back pain,and that nearly a quarter of these patients receivedtopical drugs. The drugs selected were clearly dif-ferentiated when the intensity of pain was consid-ered. As shown in Table 8, salicylates and topicaldrugs seemed to be the medication of choice forslight pain, whereas diclofenac was mainly selectedin excruciating pain. In this situation, vitamin prep-arations and topical drugs were also used (8.3%).The preferred route of administration was oral(65.2%) followed by topical (18.4) and intramuscu-lar (9.5%). However, 6.9% of patients receiveddrugs by rectal administration. Additionally, thedaily dose prescribed to each patient showed strik-ing differences when considering the distinct drugs.For example, physicians prescribed low doses ofparacetamol and aspirin in all patients. In fact, noneof them received more than 2.5 g with a mean dailyprescribed dose of 1.41 g and 1.75 g, respectively.In contrast, the majority of patients taking di-clofenac and napro.xen received established thera-peutic doses (100-200 mg for diclofenac and 500-1000 mg for naproxen).

DISCUSSION

The prevalence of pain in general population hasbeen evaluated on the basis of different methods.Using telephone calls. Crook et al. (1) found that16% of the interviewed individuals had experienced

The Clinical Journal of Pain, Vol. 6, No. 3. ¡990

270 F. BOSCH ET AL.

TABLE 7. Physician-selected therapies in the most common types of pain

SalicylatesAcetic acid derivativesPropionic acid derivativesPiroxicamParacetamolCodeine mixturesSpasmolyticsVitamin mixturesErgotic drugsTopical compoundsDigestive drugsOther drugsOther treatments

LimbsN = 133

11.322.58.36.09.00.7—3.0—

24.1—8.36.8

BackN = 111

4.514.416.23.6

14.4

——10.8—

23.4—5.47.2

Head, face,mouth

N = 62

8.16.46.4—

24.28.1—

—9.7——

33.93.2

AbdominalN = 45

—11.1

——4.4

28.9

———

28.913.313.3

ChestN = 21

14.314.34.84.84.89.5—

——14.3—

23.89.5

pain in the preceding 2 weeks. By means of a postalsurvey, Brattberg et al. (4) reported a prevalence of11% of persistent pain in Sweden. Similar ly ,Andersen and Worm-Pedersen (5) established a rateof 30% in Denmark. The present study was aimed atelucidating the characteristics rather than the prev-alence of pain complaints in patients who seek med-ical assistance for this symptom in primary healthcare in Spain. Our results confirm and extend thoseof other studies carried out in apparently healthyindividuals, reinforcing the interest of epidemio-logic surveys to evaluate the prevalence of painproblems in the community. Our present data to-gether with those previously published point out theimportance of limb and back pain in the use of med-ical care (1,3,4,6). They also demonstrate the need

to develop educational programs to improve thetraining of general practitioners in treating thesecommon complaints, because they are frequentlyrequested to face pain problems. It has been shownthat nearly 80% of individuals visit their generalpractitioner for these complaints, whereas otherhealth professionals are consulted to a much lowerextent (1). Additionally, patients with rheumatoidarthritis consider pain relief rather than reduction ofstiffness or swelling as the primary goal of treat-ment (7,8).

As indicated previously (1), women complainedof pain symptomatology more frequently than men.Men complained of musculoskeletal and traumaticpain more frequently than women, whereas the lat-ter presented headache and peripheral joint pain

TABLE 8. Intensity of pain by treatments

SalicylatesAcetic acid derivativesMetamizolArylpropionic derivativesPiroxicamOpioid drugsParacetamolCodeine mixturesSpasmolyticsVitamin mixturesErgotic drugsTopical compoundsVascular drugsDigestive drugsOther drugsOther treatments

SlightN = 24

29.24.2

———

———4.2

——25.0

——16.7

20.8

ModerateN = 131

6.99.20.85.35.3

—16.82.31.51.50.8

16.03.16.9

17.66.1

SevereN = 121

9.913.20.8

12.40.8

—13.21.71.75.01.7

14.91.73.3

10.79.1

.Very severeN = 106

0.917.04.7

11.33.83.87.5—5.76.61.9

11.31.9

—18.94.7

UnbearableN = 36

. —27.85.65.62.82.88.3

11.111.18.32.88.3

——5.6—

Results are expressed as the percentage of the intensity of pain. Drugs considered in each group are described in the footnotes ofTable 4.

The Clinical Journal of Pain. Vol. 6, No. 3. I9VO

PAIN AT PRIMAR Y CARE IN SPAIN 211

more commonly. Such incidence of higher fre-quency of headache in women had been describedearlier (3). The most frequent type of pain in men,that is, traumatic and musculoskeletal, could be at-tributed to increased loading strains at work.

An interesting feature of the present study wasthe finding that some of the so-called minor pains,such as toothache and headache, were scored in thehighest category by the patients when intensity ofpain was evaluated. These types of pain, althoughnot life threatening, may therefore be of high inten-sity when patients seek medical care.

Crook et al. (1) have shown that two-thirds ofinterviewed individuals did not visit a health pro-fessional when they had pain, mainly because thecomplaint was slight. It is therefore reasonable toassume that medical care is sought when the inten-sity of pain increases to an unbearable level inwhich self-administered therapies fail to produce re-lief. Thus, it is essential to obtain an accurate med-ication history because a considerable number ofpatients are likely to be self-treated. In this regard,Gibson and Clark showed that 14% of patients withrheumatoid arthritis were self-medicating (7). Thisvalue is probably higher, because analgesic userstend to underestimate their drug consumption(9,10). Nearly 80% of patients with rheumatoid ar-thrit is or back pain are likely to be using self-prescribed analgesics (11). This should be borne inmind in view of the report that only a half of Britishrheumatologists ask their patients about self-administered over-the-counter medication (12). Inour study, one-third of patients were taking drugsby self-prescription. Over-the-counter analgesics,such as aspirin or paracetamol, were consumed themost, but 8% used NSAIDs, mainly diclofenac andnaproxen. Additionally, a significant proportion ofpatients might have received previous treatmentsprescribed by other physicians. Taken together,these data suggest that an accurate drug history isclearly needed in order to avoid the prescription ofsimilar treatments and to allow the establishment ofcorrect dosage.

When analyzing the treatments prescribed by thephysicians participating in our study, several obser-vations deserve attention. First, the prescription ofNSAIDs for noninflammatory conditions indicatesthat these drugs might be used often for their anal-gesic effects. This would explain their high use inSpain, where NSAIDs rank in the first places ofdrug sales. A second interesting feature is the factthat diclofenac was prescribed the most to patients

with excruciating pain, suggesting that physiciansconsidered this drug very effective in these situa-tions. This belief could be explained in part by thefact that diclofenac was prescribed at therapeuticdoses, whereas aspirin and paracetamol were usedat dosages lower than those needed to obtain a fulltherapeutic effect. Alternatively, it might be arguedthat physicians had an excessive confidence on thepotency of diclofenac. The use of vitamin com-pounds to treat intolerable pain in 8% of patients isevidence of incorrect knowledge of analgesic drugson the part of primary care physicians.

In conclusion, the present study shows that limband back pain are the most frequent pain complaintsin patients who seek medical assistance at primarycare level. A high level of self-medication should beassumed in all patients presenting pain complaintsand, therefore, a careful medication history shouldbe undertaken. Finally, misconceptions about drugtherapy may explain the failure to relieve pain insome patients.

Acknowledgment: We thank all physicians who partic-ipated for their collaboration. Lluïsa Casanovas, B.Se.,and Mara Dierssen. Ph.D. provided helpful advice andcriticism.

REFERENCES

1. Crook J. Rideout E. Browne G. The prevalence of pain com-plaints in a general population. Pain 1984:18:299-314.

2. Sternbach RA. Survey of pain in the United States: the Nu-pnn Pain Report. Clin J Pain 1986:2:49-53.

3. Von Korff M. Dworkin SF, Le Resche L, Kruger A. Anepidemiologic comparison of pain complaints. Pain 1988;32:173-83.

4. Brattberg G, Thorslund M, Wikman A. The prevalence ofpain in a general population. The results of a postal survey ina county of Sweden. Pain 1989;37:215-22.

5. Andersen S. Worm-Pedersen J. The prevalence of persistentpain in a Danish population. In: Proceedings of the 5thWorld Congress on Pain. Pain 1987; (suppl 4):S332.

6. Bjelle A. Magi M Rheumatic disorders in primary care. Astudy of two primary care centres and a review of previousSwedish reports on primary care. Scand J Rheumatol 1981;10:331^*1.

7. Gibson T. Clark B. Use of simple analgesics in rheumatoidarthritis. Ann Rheum Dis 1985;44:27-9.

8. McKenna F, Wright V Pain and rheumatoid arthritis. AnnRheum Dis 1985:44:805.

9. Schwarz A. Faber U. Bomer F, Keller F, Offermann G,Molzahn M. Reliability of drug history in analgesic users.Lancet I984;2:1163-4.

10. Ready LB. Sarkis E. Turner JA. Self-reported vs actual useof medications in chronic pain patients. Pain 1982;12:285-94.

11. Higham C. Jayson M1V. Non-prescribed treatments in rheu-matic patients. Ann Rheum Dis 1982:41:203.

12. Dawes PT. Fowler PD. Shadforth MF. Hothersall TE.Availability of nonsteroidal anti-inflammatory drugs overthe counter: information needed. Br \tedJ 1984:289:413-4.

The Clinical Journal of Pain. Vol. 6. No. 3. 1990