Changes in Disability Following Physical Therapy Intervention for Patients With Low Back Pain: Dependence on Symptom Duration

Methods 

Study Populations 

Patients 

In a retrospective medical record review, we identified 133 patients seen consecutively for treatment of low back disorders between June 2002 and November 2003 at 1 of 4 University of Wisconsin Hospital and Clinics Orthopedic Outpatient sites. All subjects were classified with an International Classification of Diseases, 9th Revision14 code of 724.2 (LBP) or 724.3 (sciatica) as a primary or secondary diagnosis.

We divided patients into 3 groups based on their symptom duration. For the purpose of this study, per the parameters set by Ross,13 those beginning therapy within 1 month postonset of their symptoms were defined as the acute group, those beginning therapy between 1 month and 6 months postonset were defined as the subacute group, and those beginning therapy greater than 6 months postonset were defined as the chronic group.

The CareConnections Outcomes System (formerly TAOS) database contained information from patients who filled out the requisite initial and discharge questionnaires and had a completed episode of care. Three subjects with missing or incomplete data relating to the outcome variables were excluded from the study. Descriptive subject information concerning age, sex, and comorbidities is described in table 1. Three of the patients were receiving workers’ compensation for their back pain condition. The 3 groups were dissimilar in 3 areas: (1) the acute group was older and had a relatively higher percentage of subjects with depression, (2) the chronic group had a higher percentage of women and persons with fewer than 2 comorbidities, and (3) the subacute group had a higher percentage of persons who smoke.

Table 1. Patient Characteristics

	Characteristics	Acute (n=27)	Subacute (n=63)	Chronic (n=40)
	Age (y)
	Mean ± SD	52.9±17.0	47.3±17.4	47.1±18.5
	Range	18–82	14–87	11–81
	Sex, n (%)
	Men	12(44)	29(46)	11(27)
	Women	15(56)	34(54)	29(73)
	Depressed, n (%)
	Yes	7(26)	12(19)	7(18)
	No	20(74)	51(81)	33(82)
	Comorbidities, n (%)
	0–1	24(89)	52(83)	38(95)
	≥2	3(11)	11(17)	2(5)
	Smoking, n (%)
	Yes	2(7)	9(14)	1(3)
	No	25(93)	54(86)	39(97)

Abbreviation: SD, standard deviation.

Interventions 

Table 2 shows the percentages of patients receiving various PT interventions over the treatment duration. The most frequent combinations of these treatments are shown in table 3.

Table 2. Characteristics of Treatment Program,⁎ Utilization of Interventions, and Intensity and Duration of Care for Patients With Spinal Impairments

	Characteristics	Acute (n=27)	Subacute (n=63)	Chronic (n=40)
	Strengthening exercises	89	92	88
	Flexibility exercises	78	84	80
	Mobilization/manipulation	52	49	38
	Endurance exercises	11	10	8
	Massage techniques	22	25	20
	Heat modalities	11	10	5
	Cold modalities	7	3	8
	No. of visits
	Mean ± SD	5.5±2.9	6.8±4.4	8.0±8.0
	Range	2–12	2–32	2–51
	Treatment duration (wk)
	Mean ± SD	7.7±5.3	8.8±5.9	11.0±8.4
	Range	2–20	2–32	2–36

NOTE. Values are percentage or as otherwise indicated.

⁎ Percentage of patients receiving this type of treatment.

Table 3. Combinations of Treatments⁎ for Patients With Spinal Impairments

	Multimodal Interventions	Acute (n=27)	Subacute (n=63)	Chronic (n=40)
	Flexibility, strengthening, and mobilization/manipulation	37	40	20
	Flexibility and strengthening exercises	30	37	50
	Massage, strengthening, and mobilization/manipulation	15	14	13

NOTE. Values are percentage.

⁎ Percentage of patients receiving this combination of exercises.

The number of visits ranged from 5 to 8 for each group, over a mean duration of 7.7 weeks for the acute group, 8.8 weeks for the subacute group, and 11 weeks for the chronic group.

Results 

Mean functional outcome scores and the mean functional improvement from pretherapy to posttherapy are identified in table 4. No significant differences in initial function (P=.20) or functional outcome scores (P=.31) existed between the groups. Persons in the chronic group had significantly less functional improvement than persons in the acute group (P≤.05). Significant differences were found within each of the groups between pretreatment and posttreatment function scores. The average functional improvement score from pretreatment to posttreatment was 10.7 for the chronic group, 15.6 for the subacute group, and 19.3 for the acute group.

Table 4. Mean Assessment Measure Values

	Measure	Acute	Subacute	Chronic
	Initial function	66.9±16.7	71.9±13.4	73.1±16.8
	Functional outcome	86.3±11.2	87.5±11.6	83.7±12.9
	Functional improvement	19.3±15.1	15.6±13.4	10.7±11.7
	P	≤.05	≤.05	≤.05

NOTE. Values are mean ± SD.

The median percentage improvement scores for pain and perceived improvement are shown in table 5. The decrease in pain from pretreatment to posttreatment was significant for all groups (P<.05). There was no significant difference in percentage decrease in pain when comparing the acute, subacute, and chronic groups. However, the median percentage decrease in pain was 72.5% for the acute group, whereas the median scores for the subacute and chronic groups were only 60% and 62%, respectively. The median percentage improvement score for perceived recovery was significantly lower for the chronic group than for the acute group (P≤.05).

Table 5. Median Improvement Scores

	Outcome	Acute	Subacute	Chronic	P
	Decrease in pain⁎ (%)	72.5	60	62	NS
	Interquartile range	50–100	32–92	32–81
	Perceived improvement† (%)	87	80	74.5	.03‡
	Interquartile range	80–94	63–91	63–90

Abbreviation: NS, not significant.

⁎ Percentage decrease in pain = (initial pain − final pain)/initial pain. † Percentage improvement = perceived improvement × 10. ‡ Pairwise comparison reveals difference between acute and chronic groups

Linear regression analyses were used to identify variables with an independent association with the functional improvement score. An ANOVA of the regression line confirmed that the relation between the dependent and all independent variables is unlikely to have occurred by chance (F4=39.5, P<.001). A model with age (P=.001), symptom duration (P=.002), and inclusion of strengthening, flexibility, and mobilization/manipulation treatments (P=.001) fit the data well and explained 55.5% of the variance (table 6).

Table 6. Regression Model Explaining 55.5% of Variance in the Functional Improvement Score Among All LBP Patients

	Independent Variable	Functional Improvement Range (1–100)
		b	t	P
	Age	−0.20	−4.30	.001
	Initial function	−0.64	−11.57	.001
	Symptom duration	−3.64	−3.14	.002
	Flexibility, strengthening, and mobilization/manipulation	5.99	3.44	.001

Discussion 

The purposes of this study included assessing the impact of symptom duration on functional outcomes in patients with LBP following PT care and determine which variables were associated with enhanced function. Analysis of data showed that a combined intervention approach was effective in reducing patient pain and increasing function. The interventions selected by therapists associated with our study were not entirely consistent with results of previously published studies. Strengthening and flexibility exercises were the most frequently noted intervention choices in our study, similar to results reported by Jette and Jette8 and Jette and Delitto,10 but mobilization/manipulation techniques were utilized more frequently by our therapists. These techniques were utilized in 38% to 52% of the episodes of care (see table 2) depending on the patient acuity level, compared with a high of 39%8 and 27%10 previously reported. Some possible explanations for these discrepancies include: (1) differences in patient population or clinical environment, (2) unequal training or skill level of the therapy providers, or (3) increased research knowledge and use of evidence-based practice. Nonetheless, these choices are compliant with studies supporting its use for patients with LBP18, 19 and the inclusion of this intervention in the Guide to Physical Therapist Practice.9

Another area of discrepancy is the marked difference in utilization of heat modalities. In our study, heat was used very infrequently regardless of patient acuity (see table 2), while it was utilized for 81% and 83% to 88% of patients in the Jette and Delitto10 and Jette and Jette8 studies, respectively. The trend of adopting a more active intervention approach (eg, exercises), less reliance on passive modalities (eg, heat), and utilizing manual therapy techniques to quicken recovery is consistent with the recommended guidelines for treating patients with acute LBP.20

The change in degree of disability following PT intervention was associated with degree of patient acuity for most measures. For example, the median score for patient-perceived improvement and the functional improvement score from initial visit to discharge was significantly less for the chronic group than for the acute group. These results are consistent with those of Ross13 who found that patients with symptom duration of less than 1 month had the greatest decrease in disability and those with chronic duration (>6mo) the smallest decrease. One possible explanation for this trend is patients with longer symptom duration may have an increased possibility of developing associated psychosocial factors which could result in lower function and perceived recovery. Patients with specific psychosocial elements have been shown to have poor recovery from chronic neck and LBP at 1 year.21, 22 Our results do differ when compared with another study using CareConnections Outcomes System data, which indicated no clinically significant differences between acute, subacute, and chronic groups for patients with LBP.23 Direct comparison of results, though, is difficult due to the varying time frames marking the acute, subacute, and chronic symptom duration periods. Based on stages of tissue healing postinjury, Self et al23 used 0 to 7 days postinjury as the parameters for the acute group, 8 to 14 days for subacute, and 15 to 42 days for chronic. Considering that not all of our patients suffered an acute injury, we made the decision to use a different previously established timeframe. Last, the greatest amount of change in bodily pain, often the primary patient goal, appeared to occur in the acute phase, even though this was not statistically significant. Reduction of pain has been reported as a treatment goal for 90% of the patients treated by physical therapists for LBP.24 Of importance to clinicians is whether factors can be identified that are predictive of variability in pain and disability improvement. The importance of early intervention and rapid reduction of disability is described in studies demonstrating that in patients with LBP experiencing disability for 4 weeks or more are at higher risk for chronic disability and work restrictions.25, 26, 27, 28

Very little is known about factors that predict treatment outcome for patients with LBP. The regression analyses in this study indicated that age, symptom duration, and the inclusion of strengthening and flexibility exercises and mobilization/manipulation were the most important determinants of functional outcome score as an indication of disability. Of particular relevance, the combination of treatments that included mobilization/manipulation was the most consistent treatment predictor of better outcome. This finding differs from Jette and Jette’s report8 that inclusion of endurance exercise was most consistently associated with better outcomes, but it is consistent with recent studies that reported improved outcomes for patients with LBP and other conditions receiving mobilization/manipulation in conjunction with exercise therapy versus those patients who did not receive the manual therapy interventions.19, 29, 30 Our results support the importance of including mobilization/manipulation techniques into patient plans of care for those with LBP.

The prevalence of depression in our three patient groups ranged from 18% to 26%. Of concern is that most patients with major depression go unrecognized or are inappropriately managed.31, 32, 33 Considering this, clinicians should be screening for depression in patients with LBP. A 2-question screening instrument has been advocated for clinical use.34, 35 Whooley et al34 found that the questions: (1) “During the past month, have you often been bothered by feeling down, depressed, or hopeless?” and (2) “During the past month have you often been bothered by little interest or pleasure in doing things?” showed a 96% sensitivity (95% confidence interval [CI], 90%–99%) with a negative likelihood ratio of .07, a negative predictive value of 98%, and a specificity of 57% (95% CI, 53%–62%), with a positive likelihood ratio of 2.2 and a positive predictive value of 33%. The recommendation was that a patient “yes” response to either or both of the questions would warrant further screening, including inquiry of weight change, sleep disturbance, psychomotor problems, lack of energy, excessive guilt, poor concentration, and recurrent suicidal ideation. A cluster of 5 or more of the findings (including either or both of the initial 2 screening questions) present for at least 2 weeks causing significant distress in social, occupational, or other important areas of functioning would warrant a patient referral for further psychologic screening.36

Regarding the number of therapy visits, the acute and subacute groups averaged approximately 5 to 7 visits over 7 to 9 weeks, whereas the chronic group averaged 8 visits over 11 weeks. This study’s finding concerning visit utilization appears to be consistent with a report by Amato et al,37 who found that patients beginning therapy within the acute phase averaged fewer visits than all other groups.

Conclusions 

The aims of this clinical report were (1) to assess the impact of symptom duration on functional outcomes following rehabilitation for patients with LBP, (2) to determine the patient’s perception of recovery following discharge from therapy, and (3) to determine what variables were associated with improved function. Regardless of acuity, improvements in function and pain occur over the course of PT care, especially when strengthening, flexibility exercises, and mobilization/manipulation are included in the treatment plan. Patients seen in the chronic phase had lower functional improvement and perceived improvement scores than the acute group.