SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Objective

To describe the course of a new episode of hand and wrist problems in general practice and to identify predictors that are associated with poor outcome at short-term and long-term followup.

Methods

Patients consulting their general practitioner with hand or wrist problems (no prior consultation in the preceding 3 months) were sent a questionnaire at baseline and at 3, 6, and 12 months of followup. Potential predictors included sociodemographic variables, characteristics of the symptoms, physical activity, and psychosocial factors. General practitioners recorded information on symptoms, signs, and medical diagnosis. The main outcome measure was insufficient improvement of symptoms using the Symptom Severity Scale at short-term (3 months) and long-term (12 months) followup.

Results

Twenty-three percent of patients reported complete recovery after 3 months, increasing to 42% at 1 year after first presentation. Higher probability of poor outcome at 3 months was associated with being female, low pain intensity at baseline, and lower personal control at baseline; at 12 months it was associated with older age, being female, reporting symptoms for >3 months at baseline, low scores on the coping strategy “reducing demands,” and a higher score on somatization. Discriminative ability of the models was moderate, with areas under the curve after bootstrapping of 0.60 and 0.69 at 3 and 12 months, respectively.

Conclusion

More than half of all patients reported residual symptoms at 1 year. Although poor outcome was difficult to predict, age, sex, duration of symptoms, and psychosocial factors were associated with poor outcome of hand and wrist problems.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Musculoskeletal conditions of the hand or wrist, such as hand osteoarthritis (OA), rheumatoid arthritis (RA), or carpal tunnel syndrome (CTS), are well-recognized diagnoses in general practice. The incidence rates in general practice are estimated at 4.6 cases per 1,000 persons per year for wrist symptoms and 7.8 cases per 1,000 persons per year for hand and finger symptoms (1). The prognosis of these conditions has not yet been fully investigated in a primary care population. We know from research on other musculoskeletal disorders such as low back pain, neck pain, shoulder pain, and elbow symptoms that the intensity and course of symptoms may be influenced by sociodemographic, physical, psychological, and social factors (1–7). Information about these prognostic indicators in patients with hand or wrist problems may help general practitioners (GPs) to provide patients with adequate information regarding the most likely course of their symptoms. Such information may support decisions on treatment and referral.

In our study, we set out to study hand and wrist problems in their most general form. All types of symptoms (pain, stiffness, tingling) related to the hand or wrist were included except for symptoms caused by acute injury or vascular or skin problems. The first objective of this study was to describe the course of a new episode of hand and wrist problems in terms of perceived recovery, pain intensity, symptom severity, and perceived health. The second objective was to identify predictors that were associated with poor short-term and long-term outcome, defined as insufficient improvement of symptoms on the Symptom Severity Scale (SSS) (8). We chose to study predictors of a poor outcome rather than a good outcome, because this may help GPs to identify patients who need treatment or referral. Furthermore, identification of barriers to recovery may help to make decisions regarding the type of treatment.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Study design and recruitment.

We performed an observational study in general practice in The Netherlands. Forty-four GPs from 32 practices participated in the study. Before the start of the study, the GPs received a 3-hour instruction about the diagnosis of hand and wrist problems (history, physical examination, differential diagnosis). Between July 2004 and December 2005, the GPs recruited patients consulting for a new episode of hand or wrist problems. An episode was considered to be new if participants had not visited their GP for the same problem during the preceding 3 months. GPs asked patients to participate if they were age ≥18 years and had sufficient knowledge of the Dutch language in order to complete written questionnaires. Patients were excluded from the study if the presented symptoms were caused by an acute injury (fracture, dislocation, sprain) or by vascular or skin problems. Written informed consent was obtained from all patients.

Baseline and followup (3, 6, and 12 months) postal questionnaires were mailed to patients. Furthermore, we asked the GPs to complete a diagnosis and management registration form, on which they recorded information about history, physical examination, medical diagnoses, and management of the hand or wrist problem (wait and see, advice, splint, additional diagnostic tests, medication, and referrals). The study was approved by the Medical Ethics Committee of the VU University Medical Center in Amsterdam.

Outcome measures.

The primary outcome measure was change from baseline in symptom severity at 3 months (short-term) and 12 months (long-term) of followup, measured using the SSS (8). The SSS is a self-administered questionnaire originally developed to assess the severity of symptoms in patients with CTS. It incorporates 6 clinical areas, specifically pain, paresthesia, numbness, weakness, nocturnal symptoms, and overall functioning. The questionnaire contains 11 questions, with response options ranging from 1 (mildest) to 5 (most severe) points. The total symptom severity score is calculated as the mean of the scores for the 11 individual items. In a recent study (9), the SSS was demonstrated to be reliable and responsive in our primary care population. The minimal important change in the previous population was quantified as 0.23 points. In the present study, poor outcome was defined as a change of <0.23 points (i.e., insufficient improvement of symptom severity) at 3 and 12 months of followup, and was used as an outcome measure in the prognostic analyses.

We measured perceived recovery by asking patients if they were completely recovered from their symptoms (yes or no), and if not, they scored improvement on a 7-point transition scale (very much improved to very much deteriorated).

The third outcome measure, perceived health, was measured using the Short Form 36 health survey (SF-36) (10). The SF-36 is designed to assess 8 health concepts relevant to a person's functional status and well-being: physical functioning, role limitations in physical functioning, role limitations in emotional functioning, social functioning, bodily pain, mental health, vitality, and general health. Scale scores range from 0–100, with higher scores representing better perceived health.

Predictors of outcome.

The baseline questionnaire contained a variety of potential predictors of outcome representing sociodemographic variables and physical, psychological, and social factors. Sociodemographic factors consisted of age, sex, marital status, educational level, and work status. Body mass index (BMI) was calculated from self-reported weight and height (overweight, BMI 25–30 kg/m2; obese, BMI ≥30 kg/m2). For physical load during work and leisure time, we used the 20-item Dutch Musculoskeletal Questionnaire (0–100 scale), where 0 represents no physical workload and 100 represents highest physical workload (11).

The following characteristics of hand or wrist problems were included: duration of symptoms, previous episodes, dominant/nondominant side affected, GP diagnosis, and pain intensity (0–10-point rating scale).

For physical activity, we used 2 questions measuring frequency and intensity of physical activity. Patients were coded as meeting the Dutch Norm for Healthy Activity (yes or no) if they reported ≥30 minutes of moderate-intensity physical activity on ≥5 days of the week (12, 13). Additionally, patients were coded as meeting the American College of Sports Medicine position stand (yes or no) if they performed heavy physical exercise or sports ≥3 times a week (14).

We also assessed several psychological factors. Coping was measured with the Pain Coping Inventory consisting of 6 scales, including pain transformation, distraction, reducing demands, retreating, worrying, and resting, with a higher score indicating more use of the strategy concerned (15, 16). Personal control was measured using the personal control subscale of the Revised Illness Perception Questionnaire (1–5 scale), where a higher score indicates stronger personal control (17, 18). Distress and somatization were measured using the two 16-item subscales of the 4-Dimensional Symptom Questionnaire (0–32 scale) (19), where a cutoff score >10 for both distress and somatization discriminates between cases and noncases (20, 21). Fear-avoidance beliefs were measured using the 4-item physical activity subscale of the Fear-Avoidance Beliefs Questionnaire (0–24 scale), where a higher score indicates more fear avoidance (22). Anxiety and depressive symptoms were measured using the Hospital Anxiety and Depression Scale (HADS; 0–21 scale), where higher scores indicate more severe symptoms (23). For both subscales of the HADS, scores of 0–7 points indicate no anxiety or depression and scores of ≥8 points indicate possible or probable anxiety or depression (24). Finally, social support was measured using the Social Support Scale (12–60 scale), on which a higher score indicates less perceived support from others (25).

Statistical analysis.

Descriptive statistics were used to describe the clinical course of hand and wrist problems. A multivariate analysis of variance for repeated measures was used to test significance of changes during the 12-month followup for each outcome, and subsequently to determine at which time points changes were significant (complete-case analysis).

Univariable logistic regression analyses were performed to check whether there was a linear association between each of the potential predictors and poor outcome (i.e., less than a minimal important improvement of 0.23 points on the SSS) at 3 months and 12 months. For dichotomous variables, we only considered those variables with a prevalence of ≥10%. Potential predictors showing a nonlinear relationship with the outcome were dichotomized when cutoff scores were available from the literature. Otherwise, they were divided into tertiles (low, medium, and high), with the low category as the reference category, or when this was not possible, they were dichotomized. We calculated univariable odds ratios along with 95% confidence intervals (95% CIs). Variables that were associated with the outcome (P < 0.20) were preselected for the multivariable analysis. Before the multivariable analysis was applied, the correlation among predictors was checked. In case of a strong correlation (Spearman's r > 0.5) between variables, only the predictor with the strongest univariable association with the outcome was retained in the model.

We developed 2 multivariable models (short-term and long-term) that included the combination of predictors that was most strongly associated with poor outcome. For the short-term model, all predictors were entered simultaneously into a multivariable logistic model. However, because the number of predictors to be entered into the long-term model exceeded the number of events per 10 (events = number of patients with poor outcome), the predictors were entered in blocks (sociodemographic factors and physical factors first, characteristics of the problem next, and psychosocial factors last) (2). The best predictive model was constructed using manual backward selection. We sequentially deleted variables from the initial model until only variables with a P value less than 0.10 (Wald's statistic) were retained in the final model.

All statistical analyses were performed using SPSS for Windows, version 12.0.1 (SPSS, Chicago, IL).

Predictive performance of the models.

Calibration of the models, which is related to reliability, was assessed by plotting the predicted probabilities of poor outcome against the observed frequencies (26). For this assessment, patients were grouped into deciles according to their predicted probability. The prevalence of the outcome measure within each decile equals the observed frequency. If the predicted probabilities and the observed frequencies are in agreement, the estimates are close to the diagonal. Discrimination was studied by calculating the area under the receiver operating characteristic curve, which illustrates the ability of the models to discriminate between patients with and patients without poor outcome at subsequent cutoff points along the range of predicted probabilities (26). An area under the curve (AUC) of 0.5 indicates no discrimination above chance, whereas an AUC of 1.0 indicates perfect discrimination.

Prediction models perform better in the development cohort than in other similar populations. After the multivariable analyses, we used bootstrap samples to adjust for this over-optimism in model performance (26–28). Bootstrap samples were drawn with replacement (200 replications) from the full data set and were used to compute an adjusted AUC. This adjusted AUC provides a more precise estimation of the performance of the model in similar, future patients. The bootstrap analysis was performed using R statistical software, version 2.5.0 (R Development Core Team, Vienna, Austria).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Study population and baseline characteristics.

GPs asked 301 patients with hand or wrist problems to participate. A total of 267 (89%) patients consented to participation and completed the baseline assessment. Table 1 lists the baseline characteristics. GPs recorded >1 diagnosis in 25 (9.4%) patients. In patients given only 1 diagnosis, the 3 most frequently recorded diagnoses were OA (16.9%), tenosynovitis (15.8%), and nerve entrapment, including CTS (12.4%). The mean ± SD symptom severity score at baseline was 2.1 ± 0.6 and the mean ± SD pain intensity score was 4.0 ± 2.4; half of the patients had experienced their symptoms for >3 months when they consulted the GP.

Table 1. Patient and problem characteristics at baseline (n = 267)*
 Value
  • *

    Values are the number (percentage) unless otherwise indicated. Incidental missings (1–9).

Patient characteristics 
 Age, mean ± SD years49.3 ± 16.0
 Women198 (74.2)
 Married/living together186 (70.2)
 Education 
  Primary67 (25.2)
  Secondary141 (53.0)
  College/university58 (21.8)
 Have paid work133 (50.6)
 Body mass index, kg/m2 
  <25 (underweight/normal   weight)140 (54.3)
  25–30 (overweight)86 (33.3)
  >30 (obese)32 (12.4)
 Physical activity 
  American College of Sports Medicine position stand38 (14.6)
  Dutch Norm Healthy Activity110 (41.7)
Hand or wrist problem characteristics 
 Location of the problem 
  Unilateral198 (74.2)
  Bilateral69 (25.8)
 1 diagnosis according to general  practitioner at first  consultation 
  Osteoarthritis45 (16.9)
  Tenosynovitis42 (15.8)
  Nerve entrapment, including   carpal tunnel syndrome33 (12.4)
  Nonspecific   symptoms/unclear31 (11.7)
  Repetitive strain injury30 (11.3)
  Ganglion24 (9.0)
  Rheumatoid arthritis21 (7.9)
  Other15 (5.6)
 >1 diagnosis25 (9.4)
 Duration of symptoms at baseline 
  <2 weeks34 (12.8)
  3–4 weeks50 (18.8)
  1–2 months48 (18.0)
  3–6 months54 (20.3)
  >6 months80 (30.1)
 Severity of symptoms (1–5  scale), mean ± SD2.1 ± 0.6
 Intensity of pain (0–10 scale),  mean ± SD4.0 ± 2.4

At the first consultation, GPs prescribed medication in 36% of patients, e.g., nonsteroidal antiinflammatory drugs (n = 69) or corticosteroid injection (n = 16), and 5% were provided with splints. In 38% of patients, the treatment policy was wait and see, and 17% of patients received additional diagnostic tests, e.g., blood tests (n = 24) or radiographs (n = 22). Twenty-two percent of patients were referred, most frequently to a neurologist (n = 20), physiotherapist (n = 19), or rheumatologist (n = 6).

In total, 248 (93%) patients completed the 3-month followup questionnaire, 249 (93%) patients completed the 6-month followup questionnaire, and 248 (93%) patients completed the 12-month followup questionnaire. A total of 237 patients completed all 4 questionnaires. Baseline characteristics (including age, sex, and duration or severity of symptoms) were largely similar between these completers and the 267 enrolled patients; for example, the mean ± SD age of completers was 50.1 ± 15.9 years compared with 49.3 ± 16.0 years for noncompleters, and 75% of the completers were women compared with 74% of the noncompleters.

Clinical course.

The rates of complete recovery after 3, 6, and 12 months were 23% (n = 56), 32% (n = 80), and 42% (n = 103), respectively. Of the patients who did not report full recovery at 3 months (n = 191), 26% reported (very) much improvement, and 22% reported some improvement compared with baseline. These rates hardly changed at longer-term followup. The course of self-reported pain intensity, symptom severity, and perceived health was analyzed for the 237 patients who completed all 4 questionnaires. Self-reported pain intensity and symptom severity significantly improved during followup (P < 0.001), as well as scores on the subscales physical functioning, role physical functioning, role emotional functioning, bodily pain, and vitality of perceived health (P < 0.05) (Table 2). Pain intensity and bodily pain improved significantly at each followup measurement, symptom severity and role physical functioning improved significantly over the first 6 months, physical functioning and role emotional functioning between baseline and 3 months, and vitality between 6 months and 12 months (P < 0.05).

Table 2. Mean ± SD scores of self-reported pain intensity, symptom severity, and perceived health at baseline and 3, 6, and 12 months of followup (n = 237)
 Baseline3 months6 months12 monthsP*
  • *

    Multivariate analysis of variance.

Pain intensity, 0–10 scale4.0 ± 2.33.1 ± 2.62.6 ± 2.52.3 ± 2.50.01
Symptom severity, 1–5 scale2.2 ± 0.61.8 ± 0.71.7 ± 0.61.6 ± 0.60.01
Perceived health, 0–100 scale     
 Physical functioning77.9 ± 19.480.3 ± 20.281.3 ± 20.781.3 ± 21.40.01
 Role physical functioning61.3 ± 39.467.3 ± 40.573.1 ± 39.374.3 ± 38.70.01
 Role emotional functioning82.6 ± 33.176.3 ± 38.679.5 ± 38.082.2 ± 34.40.03
 Social functioning83.1 ± 19.082.3 ± 22.183.5 ± 21.585.5 ± 22.20.09
 Bodily pain59.7 ± 19.171.0 ± 21.075.4 ± 22.078.0 ± 23.00.01
 Mental health76.7 ± 16.374.5 ± 17.274.8 ± 18.176.3 ± 17.80.07
 Vitality65.3 ± 17.764.0 ± 19.065.1 ± 19.867.3 ± 18.00.01
 General health66.4 ± 19.867.5 ± 20.168.2 ± 20.668.5 ± 21.60.12

Short-term and long-term prognosis.

Table 3 shows the short-term and long-term univariable associations of potential predictors with poor outcome. Variables that showed univariable association (P ≤ 0.20) were entered into the multivariable model. Table 4 shows the variables that were included in the final short-term and long-term prediction models after backward selection. A higher probability of poor outcome at 3 months was associated with a combination of being female, a low pain intensity score, and lower personal control at baseline. A higher probability of poor outcome at 12 months of followup was associated with a combination of older age, being female, reporting symptoms for >3 months at baseline, lower scores on the coping strategy “reducing demands,” and a higher score on somatization.

Table 3. Univariable association of potential predictors with poor outcome at short-term (3 months) and long-term (12 months) followup*
 Short-term (n = 247)Long-term (n = 248)
OR (95% CI)POR (95% CI)P
  • *

    OR = odds ratio; 95% CI = 95% confidence interval; CTS = carpal tunnel syndrome; ACSM = American College of Sports Medicine; IPQ-R = Revised Illness Perception Questionnaire; 4DSQ = 4-Dimensional Symptom Questionnaire; FABQ = Fear-Avoidance Beliefs Questionnaire; SSS = Symptom Severity Scale; HADS = Hospital Anxiety and Depression Scale.

  • Incidental missings (1–8).

  • Incidental missings (1–9).

  • §

    P < 0.20.

Sociodemographic factors    
 Female vs. male1.83 (1.00–3.33)0.05§2.20 (1.15–4.21)0.02§
 Age per year1.01 (1.00–1.03)0.18§1.02 (1.01–1.04)0.01§
 Education level vs. primary    
  Secondary0.84 (0.46–1.55)0.580.88 (0.47–1.63)0.68
  College/university0.85 (0.40–1.79)0.660.75 (0.35–1.62)0.46
 Marital status vs. single/widowed1.15 (0.66–2.01)0.630.97 (0.55–1.74)0.93
 Having paid work vs. not having paid work0.74 (0.45–1.23)0.240.84 (0.50–1.42)0.52
 Body mass index vs. <25 kg/m2    
  25–30 kg/m20.74 (0.42–1.29)0.280.85 (0.47–1.52)0.58
  >30 kg/m20.68 (0.30–1.52)0.340.96 (0.41–2.25)0.93
 Heavy physical workload vs. none    
  Medium (12.1–25.0 vs. ≤12.0)1.08 (0.58–2.01)0.821.19 (0.63–2.25)0.60
  High (≥25.1 vs. ≤12.0)1.26 (0.69–2.33)0.451.00 (0.53–1.90)1.00
 Static posture or repetitive movements    
  Medium (22.2–44.4 vs. ≤22.1)0.77 (0.41–1.42)0.390.72 (0.38–1.35)0.31
  High (≥44.5 vs. ≤22.1)0.76 (0.40–1.43)0.390.54 (0.28–1.04)0.07§
 Sitting or visual display unit work    
  Medium (16.7–33.3 vs. ≤16.6)1.64 (0.87–3.10)0.13§0.90 (0.46–1.74)0.75
  High (≥33.4 vs. ≤16.6)1.05 (0.55–2.00)0.880.93 (0.48–1.80)0.83
Symptom characteristics    
 Duration of current symptom at baseline, ≥3 months vs. ≤2 months1.56 (0.94–2.58)0.08§2.81 (1.64–4.84)0.00§
 Recurrent problem, previous episodes vs. none1.37 (0.73–2.56)0.331.81 (0.96–3.39)0.07§
 Dominant side affected vs. none1.06 (0.62–1.82)0.821.48 (0.83–2.63)0.19§
 Diagnosis vs. all other diagnoses    
  Osteoarthritis1.32 (0.69–2.51)0.412.41 (1.23–4.69)0.01§
  Tenosynovitis0.67 (0.34–1.35)0.270.48 (0.22–1.06)0.07§
  Entrapment (including CTS)1.33 (0.60–2.96)0.490.68 (0.29–1.62)0.39
  Repetitive strain injury0.89 (0.40–1.96)0.761.29 (0.59–2.85)0.52
  Nonspecific0.72 (0.31–1.66)0.450.72 (0.30–1.73)0.46
 Pain intensity vs. 0–2    
  3–50.46 (0.25–0.86)0.01§0.86 (0.46–1.61)0.64
  6–100.61 (0.32–1.17)0.14§0.70 (0.35–1.38)0.30
Physical activity    
 ACSM position stand vs. not met0.60 (0.29–1.25)0.17§0.74 (0.35–1.59)0.45
 Dutch Norm Healthy Activity vs. not met0.92 (0.55–1.54)0.750.92 (0.54–1.56)0.75
Psychosocial factors    
 Coping with pain: Pain Coping Inventory    
  Pain transformation (7–8 vs. ≥6)1.20 (0.63–2.28)0.581.04 (0.52–2.07)0.91
  Pain transformation (≥9 vs. ≤6)1.21 (0.67–2.19)0.521.77 (0.96–3.27)0.07§
  Distraction (8–10 vs. ≤7)1.16 (0.61–2.20)0.661.13 (0.57–2.22)0.73
  Distraction (≥11 vs. ≤7)1.07 (0.58–1.95)0.841.38 (0.73–2.59)0.32
  Reducing demands (6 vs. ≤5)0.67 (0.35–1.29)0.240.62 (0.32–1.23)0.17§
  Reducing demands (≥7 vs. ≤5)0.93 (0.52–1.67)0.810.71 (0.39–1.29)0.26
  Retreating (8–9 vs. ≤7)1.01 (0.53–1.92)0.981.16 (0.60–2.26)0.66
  Retreating (≥10 vs. ≤7)1.02 (0.56–1.83)0.961.27 (0.68–2.35)0.45
  Worrying (13–16 vs. ≤12)0.84 (0.45–1.57)0.590.76 (0.40–1.44)0.40
  Worrying (≥17 vs. ≤12)0.53 (0.28–1.01)0.05§0.72 (0.37–1.38)0.33
  Resting (8–9 vs. ≤7)0.98 (0.53–1.83)0.961.94 (1.02–3.68)0.04§
  Resting (≥10 vs. ≤7)0.84 (0.46–1.53)0.570.76 (0.40–1.45)0.41
 Personal control: IPQ-R0.73 (0.53–0.99)0.04§  
  Medium (2.5–3.0 vs. ≤2.4)  0.64 (0.33–1.23)0.18§
  High (≥3.1 vs. ≤2.4)  0.58 (0.31–1.08)0.08§
 Distress: 4DSQ vs. no case0.95 (0.54–1.67)0.861.27 (0.72–2.27)0.41
 Somatization: 4DSQ vs. no case1.41 (0.80–2.48)0.242.38 (1.33–4.26)0.00§
 Fear-avoidance beliefs: FABQ    
  Medium (12–15 vs. ≤11)0.74 (0.40–1.36)0.330.72 (0.39–1.35)0.31
  High (≥16 vs. ≤11)1.17 (0.62–2.22)0.630.72 (0.37–1.41)0.34
 Social support: SSS    
  Medium (13–20 vs. 12)1.09 (0.60–1.99)0.770.99 (0.53–1.84)0.97
  High (≥21 vs. 12)0.87 (0.47–1.61)0.660.87 (0.46–1.63)0.65
 Anxiety: HADS vs. no anxiety1.16 (0.64–2.08)0.631.15 (0.63–2.13)0.65
 Depression: HADS vs. no depression0.86 (0.38–1.97)0.731.57 (0.67–3.68)0.30
Table 4. Multivariable association of predictors with poor outcome at short-term (3 months) and long-term (12 months) followup*
 Short-term (n = 239)Long-term (n = 242)
OR (95% CI)POR (95% CI)P
  • *

    OR = odds ratio; 95% CI = 95% confidence interval; PCI = Pain Coping Inventory; IPQ-R = Revised Illness Perception Questionnaire; 4DSQ = 4-Dimensional Symptom Questionnaire.

Sociodemographic factors    
 Female vs. male1.91 (1.01–3.64)0.052.12 (1.07–4.23)0.03
 Age per year  1.02 (1.01–1.04)0.01
Symptom characteristics    
 Duration of current symptom at baseline,   ≥3 months vs. ≤2 months  2.16 (1.20–3.89)0.01
 Pain intensity vs. 0–2    
  3–50.40 (0.21–0.76)0.01  
  6–100.60 (0.30–1.19)0.14  
Psychosocial factors    
 Coping with pain: PCI    
  Reducing demands (6 vs. ≤5)  0.49 (0.24–1.03)0.06
  Reducing demands (≥7 vs. ≤5)  0.58 (0.30–1.14)0.11
 Personal control: IPQ-R0.70 (0.51–0.97)0.03  
 Somatization: 4DSQ vs. no case  2.39 (1.26–4.54)0.01

Performance of the models.

Figures 1 and 2 show the calibration plots for both prognostic models. In both calibration plots, not all plotted points are close to the 45° line, demonstrating moderate calibration. Discrimination of both models was also considered to be moderate, with AUCs of 0.63 (95% CI 0.56–0.70) for the short-term model and 0.71 (95% CI 0.65–0.78) for the long-term model. After bootstrapping, the AUCs of the models were adjusted to 0.60 (3 months) and 0.69 (12 months).

thumbnail image

Figure 1. Calibration plot showing the observed frequencies versus the predicted probabilities for patients with poor outcome at 3 months of followup.

Download figure to PowerPoint

thumbnail image

Figure 2. Calibration plot showing the observed frequencies versus the predicted probabilities for patients with poor outcome at 12 months of followup.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Our observational followup study evaluated the clinical course of hand and wrist problems in primary care and investigated prognostic indicators of poor outcome. Twenty-three percent of the patients reported complete recovery after 3 months, and this proportion increased to 42% at 1 year after first presentation. A higher probability of poor outcome at 3 months was associated with a combination of being female, low pain intensity at baseline, and lower personal control at baseline. At 12 months, a higher probability of poor outcome was associated with a combination of older age, being female, reporting symptoms for >3 months at baseline, low scores on the coping strategy reducing demands, and a higher score on somatization.

During followup, gradual improvement occurred in terms of perceived recovery, pain intensity, symptom severity, and perceived health. After 12 months, however, a considerable percentage (58%) of patients still reported problems. This recovery rate is fairly consistent with the rate in studies of patients with other musculoskeletal problems. Full recovery after 6–12 months in neck, shoulder, and upper extremity problems in primary care is reported to range between 34% and 60% (2, 29–32). A substantial part of our population had chronic hand or wrist conditions such as OA and RA. Therefore, we did not expect to see full recovery in the large majority of our participants.

A recent review summarized evidence on prognostic factors for musculoskeletal pain in primary care (33). The review included prognostic studies on a wide variety of musculoskeletal conditions, but no primary care–based observational studies of hand pain were identified. Most factors that we found in our models to be associated with insufficient improvement of symptoms were also identified in this review. Older age was associated with poor outcome in several studies; for example, in low back pain, shoulder pain, and elbow pain. Likewise, longer pain duration at baseline and higher somatic perceptions were indicative of poor prognosis (33).

Irrespective of followup length, a poorer prognosis was found for female patients. OA was the most common single diagnosis, and is a chronic, long-term condition that is more prevalent in women. This may partly explain the prognostic value of sex, and the fact that a diagnosis of OA was not retained in our models. There are a number of studies on sex differences in musculoskeletal pain. These studies show a clear trend toward higher severity of pain reporting in women than in men (34–36). Explanations for these sex differences can be divided into 3 groups: 1) women are more willing to report pain than men, 2) women are more exposed to risk factors for pain than men (e.g., a study showed that at work, women spent more time using computers, did more repetitive movements, and reported using poorer and less comfortable equipment) (35), and 3) women are more vulnerable than men to develop musculoskeletal pain (35, 37). It is unclear which mechanism is most important, and research is needed to investigate the relative role of these sex differences.

In our short-term model, low pain intensity at baseline was associated with poor outcome, which means that many patients showed little improvement on the SSS. Patients with more pain at baseline have more room for improvement, resulting in a higher probability of reaching the threshold of a minimal important change. Although pain levels may reduce over time in patients with high baseline levels of pain, they may still have considerable pain at followup. Less pain intensity at baseline was also found to be a predictor of poor outcome in other prognostic studies using change in pain or symptoms as the main outcome (2, 38).

Higher scores on passive coping strategies have been reported to be associated with poor outcome across different pain syndromes (39–42). In our study, a lower score on the active coping strategy reducing demands was retained in the long-term model. This is interesting because active coping styles might be more susceptible to intervention. Further research may explore the causal association between active coping styles and outcome of symptoms, and explore the possibilities for intervention.

Illness perceptions may influence health outcomes such as pain or disability (43, 44). Personal control, which is one of the subscales of the Revised Illness Perception Questionnaire assessing illness representations, indicates the extent to which the patient believes their condition can be controlled (17, 18). In our study, low personal control was related to poor outcome at 3-month followup. This is in agreement with previous studies that have shown that a favorable course of illness is associated with high scores on perception of internal personal control (45, 46).

The performance of both models was considered moderate. The calibration plots (Figures 1 and 2) show that there was some deviation of predicted probabilities from the observed risk of poor outcome. Adjustment for over-optimism resulted in small reductions in the AUC, but the models could only moderately discriminate between patients with either good or poor outcome. However, the AUC scores found in our study were comparable to AUC scores in other studies (3, 4, 47, 48). One of the reasons the models did not fit extremely well could be the choice of our primary outcome measure, change in symptom severity, although this instrument was developed for hand problems and was demonstrated to be responsive in our hand and wrist primary care population (9). The heterogeneity of our population, including a variety of medical conditions and a number of mild self-limiting cases, may be another reason for the moderate performance of our models. Diagnosis had no predictive value in either our short-term or our long-term model. Poor outcome may be better predicted in those patients presenting to secondary care, who form a more homogeneous population with respect to severity of symptoms and diagnosis. Our models certainly identified relevant predictors, but further research is needed to confirm the predictive value of these factors in other populations.

To our knowledge, our study is the first prognostic study of hand and wrist problems in patients in primary care. The response to our study was high, with 89% of eligible and invited patients participating. The nonresponders were less often women and slightly younger than the responders, and showed a slightly different distribution of diagnoses, with a higher number of patients with RA and a lower number with OA. The response to followup was also high (93%). Baseline characteristics of the patients completing all questionnaires (n = 237) were similar to those of the enrolled population. Therefore, the models built on the completers are valid for our total study population. Our study addressed a large, heterogeneous population of primary care patients; therefore, reflecting wrist and hand problems as they are presented to the GP indicates good generalizability of our results in primary care. A variety of diagnoses was recorded by the GPs. This may have affected the performance of our models, but diagnosis was not retained in our multivariable models and other factors were more important in determining changes in symptom severity. It is possible that good predictive models can be developed within diagnostic groups (e.g., OA, RA, or CTS), but this would require larger cohorts.

In our study, we collected information about the management of hand and wrist problems by the GP. GPs prescribed medications in 36% of the patients and 22% of the patients were referred; those interventions may have influenced the prognosis. Nevertheless, we decided not to consider treatment as a potential predictor in the models. The prognostic models have been developed to help GPs make good decisions regarding treatment and referral, and should be based on general patient and disease characteristics. Confounding by indication cannot be avoided in observational studies; GPs will probably prescribe more intensive treatments to patients with more severe symptoms. Standardizing or randomizing treatment is the only way to avoid this, but it is not realistic in observational settings.

In conclusion, a poor outcome of hand and wrist problems in terms of insufficient improvement in symptom severity is difficult to predict. Nevertheless, some factors were shown to be significantly associated with poor outcome, including age, sex, duration of symptoms, and psychosocial factors. Further research should confirm associations between prognostic factors and outcome of hand and wrist problems and investigate possibilities for addressing modifiable predictors.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

Ms Spies-Dorgelo had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Spies-Dorgelo, van der Windt, Prins, van der Horst.

Acquisition of data. Spies-Dorgelo, van der Windt.

Analysis and interpretation of data. Spies-Dorgelo, van der Windt, Dziedzic, van der Horst.

Manuscript preparation. Spies-Dorgelo, van der Windt, Prins, Dziedzic, van der Horst.

Statistical analysis. Spies-Dorgelo.

Daily supervision. van der Windt, van der Horst.

Training of general practitioners. Prins.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES

We would like to thank all the patients and GPs who participated in the project.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  • 1
    Bot SD, van der Waal JM, Terwee CB, van der Windt DA, Schellevis FG, Bouter LM, et al. Incidence and prevalence of complaints of the neck and upper extremity in general practice. Ann Rheum Dis 2005; 64: 11823.
  • 2
    Bot SD, van der Waal JM, Terwee CB, van der Windt DA, Bouter LM, Dekker J. Course and prognosis of elbow complaints: a cohort study in general practice. Ann Rheum Dis 2005; 64: 13316.
  • 3
    Jellema P, van der Horst HE, Vlaeyen JW, Stalman WA, Bouter LM, van der Windt DA. Predictors of outcome in patients with (sub)acute low back pain differ across treatment groups. Spine 2006; 31: 1699705.
  • 4
    Kuijpers T, van der Windt DA, van der Heijden GJ, Twisk JW, Vergouwe Y, Bouter LM. A prediction rule for shoulder pain related sick leave: a prospective cohort study. BMC Musculoskelet Disord 2006; 7: 97.
  • 5
    Picavet HS, Schouten JS. Musculoskeletal pain in the Netherlands: prevalences, consequences and risk groups, the DMC(3)-study. Pain 2003; 102: 16778.
  • 6
    Thomas E, Silman AJ, Croft PR, Papageorgiou AC, Jayson MI, Macfarlane GJ. Predicting who develops chronic low back pain in primary care: a prospective study. BMJ 1999; 318: 16627.
  • 7
    Walker-Bone K, Palmer KT, Reading I, Coggon D, Cooper C. Prevalence and impact of musculoskeletal disorders of the upper limb in the general population. Arthritis Rheum 2004; 51: 64251.
  • 8
    Levine DW, Simmons BP, Koris MJ, Daltroy LH, Hohl GG, Fossel AH, et al. A self-administered questionnaire for the assessment of severity of symptoms and functional status in carpal tunnel syndrome. J Bone Joint Surg Am 1993; 75: 158592.
  • 9
    Spies-Dorgelo MN, Terwee CB, Stalman WA, van der Windt DA. Reproducibility and responsiveness of the Symptom Severity Scale and the hand and finger function subscale of the Dutch Arthritis Impact Measurement Scales (Dutch-AIMS2-HFF) in primary care patients with wrist or hand problems. Health Qual Life Outcomes 2006; 4: 87.
  • 10
    Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992; 30: 47383.
  • 11
    Bot SD, Terwee CB, van der Windt DA, Feleus A, Bierma-Zeinstra SM, Knol DL, et al. Internal consistency and validity of a new physical workload questionnaire. Occup Environ Med 2004; 61: 9806.
  • 12
    Kemper H, Ooijendijk W, Stiggelbout M. Consensus about the Dutch recommendation for physical activity to promote health. Tijdschrift voor Gezondheidswetenschappen 2000; 78: 1803.In Dutch.
  • 13
    Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C, et al. Physical activity and public health: a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 1995; 273: 4027.
  • 14
    American College of Sports Medicine position stand: the recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness in healthy adults. Med Sci Sports Exerc 1990; 22: 26574.
  • 15
    Kraaimaat F, Bakker A, Evers A. Pain coping strategies in chronic pain patients: the development of the Pain Coping Inventory (PCI). Gedragstherapie 1997; 30: 185201.In Dutch.
  • 16
    Kraaimaat FW, Evers AW. Pain-coping strategies in chronic pain patients: psychometric characteristics of the Pain-Coping Inventory (PCI). Int J Behav Med 2003; 10: 34363.
  • 17
    Moss-Morris R, Weinman J, Petrie KJ, Horne R, Cameron LD, Buick D. The revised Illness Perception Questionnaire (IPQ-R). Psychol Health 2002; 17: 116.
  • 18
    Weinman J, Petrie KJ, Moss-Morris R, Horne R. The Illness Perception Questionnaire: a new method for assessing the cognitive representation of illness. Psychol Health 1996; 11: 43145.
  • 19
    Terluin B, van Rhenen W, Schaufeli WB, de Haan M. The Four-Dimensional Symptom Questionnaire (4DSQ): measuring distress and other mental health problems in a working population. Work Stress 2004; 18: 187207.
  • 20
    Terluin B. The Four Dimensional Symptom Questionnaire (4DSQ) in general practice. De Psycholoog 1998; 33: 1824.In Dutch.
  • 21
    Terluin B, van Marwijk HW, Ader HJ, de Vet HC, Penninx BW, Hermens ML, et al. The Four-Dimensional Symptom Questionnaire (4DSQ): a validation study of a multidimensional self-report questionnaire to assess distress, depression, anxiety and somatization. BMC Psychiatry 2006; 6: 34.
  • 22
    Waddell G, Newton M, Henderson I, Somerville D, Main CJ. A Fear-Avoidance Beliefs Questionnaire (FABQ) and the role of fear-avoidance beliefs in chronic low back pain and disability. Pain 1993; 52: 15768.
  • 23
    Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983; 67: 36170.
  • 24
    Bjelland I, Dahl AA, Haug TT, Neckelmann D. The validity of the Hospital Anxiety and Depression Scale: an updated literature review. J Psychosom Res 2002; 52: 6977.
  • 25
    Feij J, Doorn C, van Kampen D, van den Berg P, Resing W. Sensation seeking and social support as moderators of the relationship between life events and physical illness/psychological distress. In: WinnubstJA, MaesS, editors. Lifestyles stress and health. Leiden: DSWO Press; 1992. p. 285302.
  • 26
    Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med 1996; 15: 36187.
  • 27
    Efron B, Tibshirani R. An introduction to the bootstrap: monographs on statistics and applied probability. New York: Chapman and Hall; 1993.
  • 28
    Steyerberg EW, Harrell FE Jr, Borsboom GJ, Eijkemans MJ, Vergouwe Y, Habbema JD. Internal validation of predictive models: efficiency of some procedures for logistic regression analysis. J Clin Epidemiol 2001; 54: 77481.
  • 29
    Croft P, Pope D, Silman A, for the Primary Care Rheumatology Society Shoulder Study Group. The clinical course of shoulder pain: prospective cohort study in primary care. BMJ 1996; 313: 6012.
  • 30
    Feleus A, Bierma-Zeinstra SM, Miedema HS, Verhagen AP, Nauta AP, Burdorf A, et al. Prognostic indicators for non-recovery of non-traumatic complaints at arm, neck and shoulder in general practice: 6 months follow-up. Rheumatology (Oxford) 2007; 46: 16976.
  • 31
    Van der Windt DA, Koes BW, Boeke AJ, Deville W, de Jong BA, Bouter LM. Shoulder disorders in general practice: prognostic indicators of outcome. Br J Gen Pract 1996; 46: 51923.
  • 32
    Winters JC, Sobel JS, Groenier KH, Arendzen JH, Meyboom-de Jong B. The long-term course of shoulder complaints: a prospective study in general practice. Rheumatology (Oxford) 1999; 38: 1603.
  • 33
    Mallen CD, Peat G, Thomas E, Dunn KM, Croft PR. Prognostic factors for musculoskeletal pain in primary care: a systematic review. Br J Gen Pract 2007; 57: 65561.
  • 34
    Tsai YF. Gender differences in pain and depressive tendency among Chinese elders with knee osteoarthritis. Pain 2007; 130: 18894.
  • 35
    Wijnhoven HA, de Vet HC, Picavet HS. Explaining sex differences in chronic musculoskeletal pain in a general population. Pain 2006; 124: 15866.
  • 36
    Wijnhoven HA, de Vet HC, Picavet HS. Prevalence of musculoskeletal disorders is systematically higher in women than in men. Clin J Pain 2006; 22: 71724.
  • 37
    Strazdins L, Bammer G. Women, work and musculoskeletal health. Soc Sci Med 2004; 58: 9971005.
  • 38
    Bot SD, van der Waal JM, Terwee CB, van der Windt DA, Scholten RJ, Bouter LM, et al. Predictors of outcome in neck and shoulder symptoms: a cohort study in general practice. Spine 2005; 30: E45970.
  • 39
    Burton AK, Tillotson KM, Main CJ, Hollis S. Psychosocial predictors of outcome in acute and subchronic low back trouble. Spine 1995; 20: 7228.
  • 40
    Covic T, Adamson B, Hough M. The impact of passive coping on rheumatoid arthritis pain. Rheumatology (Oxford) 2000; 39: 102730.
  • 41
    Evers AW, Kraaimaat FW, Geenen R, Bijlsma JW. Psychosocial predictors of functional change in recently diagnosed rheumatoid arthritis patients. Behav Res Ther 1998; 36: 17993.
  • 42
    Evers AW, Kraaimaat FW, Geenen R, Jacobs JW, Bijlsma JW. Pain coping and social support as predictors of long-term functional disability and pain in early rheumatoid arthritis. Behav Res Ther 2003; 41: 1295310.
  • 43
    Harkapaa K, Jarvikoski A, Estlander AM. Health optimism and control beliefs as predictors for treatment outcome of a multimodal back treatment program. Psychol Health 1996; 12: 12334.
  • 44
    Heijmans M, de Ridder D. Assessing illness representations of chronic illness: explorations of their disease-specific nature. J Behav Med 1998; 21: 485503.
  • 45
    Hagger MS, Orbell S. A meta-analytic review of the common-sense model of illness representations. Psychol Health 2003; 18: 14184.
  • 46
    Scharloo M, Kaptein AA, Weinman J, Hazes JM, Willems LN, Bergman W, et al. Illness perceptions, coping and functioning in patients with rheumatoid arthritis, chronic obstructive pulmonary disease and psoriasis. J Psychosom Res 1998; 44: 57385.
  • 47
    Kuijpers T, van der Windt DA, Boeke AJ, Twisk JW, Vergouwe Y, Bouter LM, et al. Clinical prediction rules for the prognosis of shoulder pain in general practice. Pain 2006; 120: 27685.
  • 48
    Schoonhoven L, Grobbee DE, Donders AR, Algra A, Grypdonck MH, Bousema MT, et al, for the prePURSE Study Group. Prediction of pressure ulcer development in hospitalized patients: a tool for risk assessment. Qual Saf Health Care 2006; 15: 6570.