Evaluation of Working Health Services Scotland, 2010-2014

4 Discharge

4.1 Cases that completed the programme

As discussed in Section 2.3, two indications can be used for whether cases completed the programme; whether they had any relevant data in the discharge paperwork (N=7,022); and whether they had completed the EQ-5D at discharge (which is a subset of the first group, with a more complete set of discharge paperwork, N=5,590).

Overall, half (50.3%) of the 11,103 cases, for whom entry EQ-5D scores are recorded, completed the discharge EQ-5D paperwork. This proportion is maintained across gender, primary health condition, SIMD and absence status at entry (Figure 21). Differences in recorded completion rate were seen when analysed by age, with the proportion who completed the programme being 36% among the youngest (<30) and 55% for the oldest age group (50+). Differences in completion by Board area are discussed in Section 3.4.

Figure 21: Percentage of cases who completed the programme compared to those who entered WHSS (completed the entry assessment)

4.2 Changes in absence status while in programme

The majority of cases (75%) were at work both at entry and discharge from the programme, while 4% were off work at entry and discharge. However, 18% (1,188 cases) who were absent when they entered the programme were at work on discharge from it. Two percent were at work at entry and on sick leave at discharge (Table 11).

Table 11: Showing the change in absence status from entry to discharge

		Absence status at discharge		Total
		At work	Off sick
Absence status at entry	At work	4,933 (75%)	154 (2%)	5,087 (78%)
	Off sick	1,188 (18%)	266 (4%)	1,454 (22%)
	Total	6,121 (94%)	420 (6%)	6,541 (100%)

The analysis in this section is based on the response to the question at discharge "Are you currently off sick?" ^[4] , for the cases for whom data are available at entry as well as discharge. Altogether, 94% of the 6,541 cases were at work at discharge and 6% were absent. Significant differences between groups are:

• The cases in the most deprived group ( SIMD 1) are more likely to be on sick leave at discharge compared to the cases in the least deprived group ( SIMD 5) ( RR = 1.89).
• Cases who are aged 50+ are more likely to be on sick leave at discharge compared to 30-39 years old group ( RR = 1.44).
• The cases with MH conditions are more likely to be on sick leave at discharge than those with MSK conditions ( RR = 1.89).
• A case who is off sick at entry is 6 times more likely to be off sick at discharge compared to a case who is at work at entry ( RR = 6.04).
• A case who is at work at entry is 1.2 times more likely to be at work at discharge compared to a case who is off sick at entry ( RR = 1.19).

4.3 Number of lost working days during the programme

Cases were asked how many working days they had lost due to sickness absence for their primary health condition while in the programme (Table 12). Note that this is self-reported, and only those who reported losing some time during the programme are shown.

Table 12: Average number of working days lost while in WHSS, by absence status at entry

Entry	Discharge	N	Average working days lost while in WHSS	Standard Deviation (working days)
At work	At work	489	10.0	17.3
	Absent	48	38.5	51.9
	Unknown	59	24.8	39.4
Absent	At work	616	28.0	34.0
	Absent	123	66.0	63.3
	Unknown	96	71.1	64.4
Total		1,431	28.2	41.6

A return to work date was provided by 551 cases who were absent at entry; it was therefore possible to calculate the number of calendar days from their enrolment assessment to their return to work (i.e. their absence while in the programme). They had an average absence while in the programme of 51.8 calendar days ( SD = 63.8 calendar days). This same group self-reported that they had lost an average of 31.0 working days while in the programme ( SD = 37.1 working days).

For the 1,431 cases who provided information on the number of working days lost while in the programme, the average number of working days lost was 28.2 ( SD=41.6). For the 1,063 MSK cases, the self-reported number of working days lost while in WHSS was 22.5 ( SD=33.1); while it was twice as long for the 291 MH cases, at 44.0 working days ( SD=54.5).

4.4 Association between lost working time while in the programme and other factors

The results of the time series analytical model on the duration of sickness absence indicate that age, length of absence prior to entering the programme, primary condition and duration in the programme are related to the number of lost days while in the programme; gender and SIMD did not significant influence the number of lost days while in the programme. The autoregressive model suggests that older cases took longer to return to work, with almost 5 more days for every 10 additional years of age. The number of days lost due to sickness absence while in the programme is significantly higher in MH cases (p<0.001) than MSK cases; 50% of MSK cases are back to work in 21 days, while this is 46 days for MH cases (Figure 22).

In general, those who had been off sick for a longer time prior to entry took longer to go back to work during the intervention. The time it took for 50% of the cases who had been on sick leave for up to 2 weeks prior to entering the programme to return to work was 23 days, while it was 63 days for those who had been off sick for 9 to 11 weeks. There is a significant difference between the cases who were off sick for the shortest time before entry assessment (0-2 weeks) and those who were off sick for over 9 weeks (p<0.0001) (Figure 23).

Figure 22: Kaplan Meier return to work curve by primary condition
[ MH=Mental health cases; MSK=musculoskeletal cases]

Figure 23: Median length of time (days) being off sick since assessment by the number of weeks cases have been off sick before assessment.

The best statistical model to model duration of sickness absence for the cases that were off sick at entry assessment includes age, length of time between sick leave and entering the programme, primary health condition and discharge time. Gender (p=0.32), SIMD (p=0.19), occupation category (p=0.35), and general health status at entry assessment ( EQ-5D index) (p=0.36) were not significant factors and therefore are not included in the model. HADS and COPM scores were not included in this analysis, as this would have reduced the sample size and thus the analysis power of the model.

As seen in Table 13, age is positively correlated to the duration of sickness absence and adds almost 5 days to the duration of absence for every 10 year age category i.e. older cases took longer to return to work, with almost 5 more days of absence for every 10 additional years of age. The referrals that had longer sickness absence prior to entering the programme also had a longer sickness absence during the programme. Duration of sickness absence prior to entry assessment was re-coded for the analysis (1= 0 to 2 weeks; 2= 3 to 5 weeks; 3= 6 to 8 weeks; 4= 9 to 11 weeks; 5=over 12 weeks). By moving up in sickness absence prior to entry categories, 10 days are added to the duration of sickness i.e. the model suggests that those who were absent for less than 2 weeks prior to entry to the programme, had absences of almost 10 days less than those who had been off for 3-5 weeks prior to entry, and almost 40 days less than those who had been off for over 12 weeks prior to entry. The impact of primary health condition was analysed using mental health as the reference category. The statistical model also suggests that those who presented with MSK conditions had 10 days less sickness absence while in the programme than those with MH conditions. Longer periods of sickness absence while in WHSS were associated with longer durations in the programme; for every 10 additional days in cases' discharge time from the programme, their sickness absence duration while in the programme increased by 2 days.

Table 13: Model result of ARIMA model for duration of sickness absence of WHSS referrals

Model parameters	Estimate (days)	Standard error of estimates	Z- statistic	P-value
Age (years)	0.49	0.16	3.04	0.003
Duration of sickness absence prior to entering the programme (ref= 0-2 weeks)	9.91	1.81	5.47	<0.0001
Primary condition (ref= MH)	-10.60	3.65	2.90	0.006
Discharge time (days)	0.22	0.03	8.23	<0.0001

4.5 Health issue resolved

Of the 7,869 cases who responded to the question at discharge, on whether the health issue was resolved, 77% answered positively (34% fully resolved, 43% partly resolved). The proportion saying that their health issue had resolved was lowest in SIMD 1 (74%) and highest in SIMD 5 (81%).

Considering this by primary health condition, 80% of those who had a MSK condition at entry considered that their health condition was either fully or partially resolved at discharge, while 83% of the MH cases considered the same (Figure 24).

Figure 24: The percentage of MSK and MH cases who thought their health condition had resolved at discharge

Considering this by absence status at entry, 81% who were at work at entry (N=5,784) reported that their health condition was either partly or fully resolved at discharge, while this was 74% for those who were absent at entry (N=1,863) (Figure 25).

Figure 25: The percentage of those who were absent / at work at entry who thought their health condition had resolved at discharge

4.6 Changes in health tool scores at discharge

4.6.1 Overview

Entry and discharge health measures were statistically compared for the cases that completed the programme, for which the EQ-5D scores were available at entry and discharge. The HADS and COPM scores at entry and discharge were analysed where these data were available from the group for whom there were completed EQ-5D at entry and discharge. This is summarised in Table 14.

Table 14: Average changes in health measure scores

Measure	Pre intervention mean score	Post-intervention mean score	Average change in score	Number
EQ-5D index
All completers	0.51	0.81	0.30	5,590
MSK cases	0.50	0.81	0.31	4,749
MH cases	0.58	0.84	0.26	646
EQ-5D VAS score
All completers	59.1	80.0	22.5	5,472
MSK cases	60.6	80.8	22.5	4,653
MH cases	48.8	76.2	30.0	631
COPM Performance score
All completers	3.84	7.54	3.70	3,771
MSK cases	3.91	7.62	3.71	3,182
MH cases	3.27	7.26	3.99	457
COPM Satisfaction score
All completers	2.87	7.44	5.00	3,754
MSK cases	2.91	7.53	5.00	3,166
MH cases	2.46	7.18	5.00	457
HADS anxiety score
All completers	7.36	4.04	-3.32	1,696
MSK cases	5.57	3.26	-2.31	1,203
MH cases	12.67	6.18	-6.50	400
HADS depression score
All completers	5.94	2.80	-3.14	1,696
MSK cases	4.68	2.33	-2.35	1,203
MH cases	9.65	3.98	-5.67	400

Table 14 shows the mean entry assessment (pre-intervention) and discharge (post-intervention) scores for the EQ-5D index value; the EQ-5D VAS, COPM Performance and Satisfaction scores; and HADS Anxiety and Depression scores. The average change in score is also shown. For all health measures the changes from entry to discharge are statistically significant (p<0.001), and this remains the case when the changes are considered by primary health condition ( MSK and MH). Note that the figures for 'All completers' are similar to the figures for ' MSK cases' as the 'All completers' population is largely made up of MSK cases (approximately 84%).

Note also that a negative change in score for the HADS anxiety and depression scores indicates an improvement. Not surprisingly, the change in HADS scores for cases with MH primary condition is greater than for cases with MSK primary condition (as it is measuring anxiety and depression), while the changes in the other health measures are more similar when comparing the two health conditions.

A multivariate logistic regression model suggests that the number of services offered to the cases influences the odds of completing the discharge paperwork (and therefore the programme); the odds of completing the discharge are reduced if more services are offered to the cases. Also those cases whose interventions lasted for longer periods were less likely to complete the discharge. The number of services offered and the duration of use of these services are likely to indicate a complex health need, which could be a reason for non-completion of the programme.

4.6.2 Changes in EQ-5D index scores

Changes in EQ-5D index scores were calculated for the 5,590 cases for whom there are entry and discharge scores. The changes in EQ-5D index range from -0.92 to 1.41 (negative sign means the health got worse from entry to discharge). The average change in score is 0.30, which is statistically significant from zero (p<0.001). Altogether, 4,920 cases (88%) improved their index score (by an average of 0.35 points); 5% of cases did not change their index score; 7% had a worse index score (by an average of 0.15 points). The extent of the positive change is striking from a health economic perspective, and although there is no control group, it cannot be ruled out that the WHSS intervention has contributed to this health benefit.

When considering the change in EQ-5D index score by primary health condition, Figure 26 shows that 89% of MSK cases and 84% of MH cases improved their EQ-5D score, with a slightly greater increase in score for the MSK cases (0.36 compared with 0.33 for MH cases). Altogether, 6% of MSK cases and 9% of MH cases had a worse score at discharge, with similar values (-0.15 and -0.13 respectively).

The mean score of MSK cases was 0.50 at entry which rose to 0.81 at discharge (N=4,749). For MH cases the mean scores rose from 0.58 to 0.84 (N=646).

Figure 26: Change in EQ-5D index values shown for MSK and MH cases

The EQ-5D index value for those at work and absent at entry is shown in Figure 27.

Figure 27: Change in EQ-5D index values shown for those at work / absent at entry

Figure 27 shows that 88% both of those who were at work and those who were absent at entry had a better EQ-5D index score at discharge. The improvement was greater for those who were absent at entry (0.42 compared with 0.33 for those at work at entry). Altogether, 7% of those who were at work at entry and 8% of those absent at entry had a worse score at discharge, with those who were absent at entry having slightly worse scores (-0.18 compared to -0.14).

The mean score of those who were at work when they entered the programme was 0.54 which rose to 0.82 at discharge (N=4,267). For those who were absent at entry, their mean scores rose from 0.42 to 0.77 (N=1,323).

4.6.3 Change in EQ-5D Visual Analogue Scale ( VAS)

Change in VAS scores were calculated for 5,472 cases whose entry and discharge VAS scores were available. Altogether 4,429 cases (81%) improved their score. The average change in scores was 21 points, which is significantly above zero (p<0.001).

When considering this by primary health condition, Figure 28 shows that 81% of MSK cases and 87% of MH cases improved their EQ-5D score, with a greater increase in score for the MH cases of 6 points (33 compared with 27 for MSK cases). Altogether, 10% of MSK cases and 6% of MH cases had a worse score at discharge, with average values of -16 and -20 respectively.

The mean VAS score for MSK cases was higher at entry (61) and discharge (81)[N=4,653] than for MH cases, being 49 at entry and 76 at discharge [N=631].

Figure 28: Change in EQ-5D VAS scores shown for MSK and MH cases

When considering EQ-5D VAS score by absence status when entering the programme, Figure 29 shows that 81% of those who were at work and 82% of those who were absent at entry had a better EQ-5D index score at discharge. The improvement was greater for those who were absent at entry (32 points compared with 26 points for those at work at entry). Altogether, 10% of both those who were at work and those absent at entry had a worse score at discharge, with those who were absent at entry having a worse average score (-21) compared to those who were at work at entry (-16).

The mean score of cases who were at work when they entered the programme was 62.6 which rose to 81.4 at discharge (N=4,189), an increase of 19.8 points. The mean scores of cases who were absent at entry rose from 51.0 to 75.5 (N=1,283), an increase of 24.5 points.

Figure 29: Change in EQ-5D VAS scores shown for those at work / absent at entry

4.6.4 Overview of HADS scores

HADS anxiety and depression scores are available for 1,696 of the cases who completed the EQ-5D at discharge; 71% of these are cases with MSK as a primary condition and 24% have an MH condition as a primary condition, while 5% have an 'other' health condition. HADS scores range from 0 to 21, while 0-7 is considered normal, 8-10 borderline and 11-21 is 'caseness'. The following sections present the changes in caseness status and in HADS score for the anxiety scores (section 4.6.5) and the depression scores (section 4.6.6).

4.6.5 HADS anxiety scores

The status relating to anxiety of all cases in pre- and post-intervention for the three categories is given in Table 15.

Table 15: HADS anxiety status at pre- and post-intervention

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	926 (54.6%)	29 (1.7%)	13 (0.8%)	968 (57.1%)
	Borderline	182 (10.7%)	35 (2.1%)	12 (0.7%)	229 (13.5%)
	Caseness	288 (17.0%)	104 (6.1%)	107 (6.3%)	499 (29.4%)
Total		1,396 (82.3%)	168 (9.9%)	132 (7.8%)	1,696 (100%)

Table 15 shows that 29% of cases had 'caseness' anxiety status at entry to the programme while only 8% did at discharge. In addition, 574 cases (34%) have moved to a healthier HADS anxiety category, among which 288 cases (17%) changed from caseness to normal. However, 54 (3%) moved to a poorer HADS anxiety category. Altogether, 55% were considered as normal and did not change during intervention.

The changes in HADS anxiety status for the MSK cases are shown in Table 16 and for MH cases in Table 17. This shows differences between the groups, reflecting the fact that the HADS measures the mental health issues of anxiety and depression. Altogether 23% of MSK cases have moved to a healthier HADS anxiety score, while 69% of MH cases have. A table is not shown for those with an 'other' primary health condition.

Table 16: HADS anxiety status at pre- and post-intervention for MSK cases

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	820 (68.2%)	23 (1.9%)	11 (0.9%)	854 (71.0%)
	Borderline	125 (10.4%)	22 (1.8%)	9 (0.7%)	156 (13.0%)
	Caseness	101 (8.4%)	48 (4.0%)	44 (3.7%)	193 (16.0%)
Total		1,046 (86.9%)	93 (7.7%)	64 (5.3%)	1,203 (100%)

Table 17: HADS anxiety status at pre- and post-intervention for MH cases

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	56 (14.0%)	3 (0.8%)	1 (0.3%)	60 (15.0%)
	Borderline	47 (11.8%)	10 (2.5%)	2 (0.5%)	59 (14.8%)
	Caseness	175 (43.8%)	52 (13.0%)	54 (13.5%)	281 (70.3%)
Total		278 (69.5%)	65 (16.3%)	57 (14.3%)	400 (100%)

Table 18 and 19 show these changes in score for cases who were at work at entry (Table 18) and those who were absent at entry (Table 19). Altogether 29% of cases who were at work at entry have moved to a healthier HADS anxiety score, while 46% of those who were absent have.

Table 18: HADS anxiety status at pre- and post-intervention for those at work at entry

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	746 (61.9%)	23 (1.9%)	8 (0.7%)	777 (64.4%)
	Borderline	123 (10.2%)	28 (2.3%)	2 (0.2%)	153 (12.7%)
	Caseness	164 (13.5%)	61 (5.1%)	52 (4.3%)	276 (22.9%)
Total		1,032 (85.6%)	112 (9.3%)	62 (5.1%)	1,206 (100%)

Table 19: HADS anxiety status at pre- and post-intervention for those absent at entry

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	180 (36.7%)	6 (1.2%)	5 (1.0%)	191 (39.0%)
	Borderline	59 (12.0%)	7 (1.4%)	10 (2.0%)	76 (15.5%)
	Caseness	125 (25.5%)	43 (8.8%)	55 (11.2%)	223 (45.5%)
Total		364 (74.3%)	56 (11.4%)	70 (14.3%)	490 (100%)

Regarding the HADS Anxiety score (rather than health category - normal, borderline or caseness), 1,235 (73%) cases improved their scores, with an average change of 5.1 ^[5] . The change was more marked for the MH cases (89% improved, with an average improvement of 7.5 points, N=400) than the MSK cases (67% improved, with an average change of 4.1, N=1,203). Altogether, 7% of MH cases had a worse score, with an average of 2.4; while 16% of MSK cases had a worse score, with an average of 2.9. The score was unchanged for 4% of MH cases and 16% of MSK cases.

Of those who were off sick at entry, 80% improved their HADS Anxiety scores by an average of 5.9 points; 14% had a worse score (average of 3.3) while 7% had the same score (N=490). For those at work at entry, 70% improved their score by an average of 4.7 points; 14% had a worse score (average of 2.7 points) while 16% had the same score (N=1,206).

4.6.6 HADS depression scores

The HADS depression health category status of all cases at pre- and post-intervention is given in Table 20.

Table 20: HADS depression status at pre- and post-intervention

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	1,100 (64.9%)	28 (1.7%)	13 (0.8%)	1,141 (67.3%)
	Borderline	211 (12.4%)	27 (1.6%)	11 (0.6%)	249 (14.7%)
	Caseness	209 (12.3%)	35 (2.1%)	62 (3.7%)	306 (18.0%)
Total		1,520 (89.6%)	90 (5.3%)	86 (5.1%)	1,696 (100%)

In Table 20, 18% of cases had a 'caseness' depression status at entry, which dropped to 5% at discharge. Furthermore, 455 (27%) of cases have moved to a healthier HADS depression category, among which 209 cases (12%) have moved from caseness to normal. However, 52 cases (3%) have moved to a poorer HADS depression category. Altogether, 65% were considered as normal at entry, and did not change during the intervention.

Table 21 and 22 show these changes in score for cases with MSK (Table 21) and MH as a primary health condition (Table 22). Altogether 17% of MSK cases have moved to a healthier HADS depression score, while 55% of MH cases have.

Table 21: HADS depression status at pre- and post-intervention for MSK cases

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	916 (76.1%)	19 (1.6%)	6 (0.5%)	941 (78.2%)
	Borderline	113 (9.4%)	17 (1.4%)	6 (0.5%)	136 (11.3%)
	Caseness	82 (6.8%)	14 (1.2%)	30 (2.5%)	126 (10.5%)
Total		1,111 (92.4%)	50 (4.2%)	42 (3.5%)	1,203 (100%)

Table 22: HADS depression status at pre- and post-intervention for MH cases

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	127 (31.8%)	7 (1.8%)	5 (1.3%)	139 (34.8%)
	Borderline	85 (21.3%)	8 (2.0%)	3 (0.8%)	96 (24.0%)
	Caseness	119 (29.8%)	17 (4.3%)	29 (7.3%)	165 (41.3%)
Total		331 (82.8%)	32 (8.0%)	37 (9.3%)	400 (100%)

Table 23 and 24 show these changes in score for cases who were at work at entry (Table 23) and those who were absent at entry (Table 24). Altogether 20% of cases who were at work at entry have moved to a healthier HADS depression score, while 43% of those who were absent have.

Table 23: HADS depression status at pre- and post-intervention for those who were at work at entry

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	892 (74.0%)	19 (1.6%)	6 (0.5%)	917 (76.0%)
	Borderline	124 (10.3%)	18 (1.5%)	6 (0.5%)	148 (12.3%)
	Caseness	105 (8.7%)	14 (1.2%)	22 (1.8%)	141 (11.7%)
Total		1,121 (93.0%)	51 (4.2%)	34 (2.8%)	1,206 (100%)

Table 24: HADS depression status at pre- and post-intervention for those who were absent at entry

		Post-intervention			Total
		Normal	Borderline	Caseness
Pre-intervention	Normal	208 (42.4%)	9 (1.8%)	7 (1.4%)	224 (45.7%)
	Borderline	87 (17.8%)	9 (1.8%)	5 (1.0%)	101 (20.6%)
	Caseness	104 (21.2%)	21 (4.3%)	40 (8.2%)	165 (33.7%)
Total		399 (81.4%)	39 (8.0%)	52 (10.6%)	490 (100%)

Altogether, 1,226 cases (72%) improved their HADS Depression scores, with an average change of 4.9. As with the HADS Anxiety scores, the change was more marked for the MH cases (85% improved, with an average improvement of 7.0 points, N=400) than the MSK cases (68% improved, with an average change of 4.0, N=1,203). Altogether, 10% of MH cases had a worse score, with an average of 3.0; while 15% of MSK cases had a worse score, with an average of 2.5. The score was unchanged for 5% of MH cases and 16% of MSK cases.

Of those who were off sick at entry, 79% improved their HADS Depression scores by an average of 6.1 points; 11% had a worse score (average of 3.4) while 10% had the same score (N=490). For those at work at entry, 69% improved their score by an average of 4.3 points; 15% had a worse score (average of 2.5 points) while 15% had the same score (N=1,206).

4.6.7 Change in COPM scores

COPM Performance and Satisfaction scores range from 0 to 10; a higher score represents better performance, and better satisfaction with performance. COPM scores are considered for those for whom EQ-5D scores are available at entry and discharge.

Altogether, 89% of cases improved their COPM Performance score (N=3,771), with the average change score being 4.2 (statistically significant). Only 3% had a worse COPM Performance score (average of 1.7). The improvements were similar for both MSK and MH cases: 89% of MSK cases (N=3,182) improved their score by an average of 4.2, and 93% of MH cases (N=457) improved their score by an average of 4.3 (see Figure 30). The mean score at entry changed from 3.9 to 7.6 at discharge for MSK cases, while from 3.3 to 7.3 for MH cases.

Figure 30: Change in COPM Performance scores shown for MSK and MH cases

The percentage that improved was also similar when comparing those who were at work at entry (89%, N=2,843) with those who were absent (90%, N=928), see Figure 31. However the size of the change of score was greater for those who were absent at entry, being an average of 5.0, compared with 3.9 for those who were at work at entry. The entry scores were lower for those who were absent at entry (2.8) and were 7.3 at discharge, while the entry scores were higher for those who were at work at entry (4.2) and were also slightly higher at discharge (7.6).

Figure 31: Change in COPM Performance scores shown for those at work / absent at entry

Similarly, 90% of cases improved their COPM Satisfaction score (N=3,754), with the average change in scores being 5.1 (statistically significant). Only 3% had a worse COPM Satisfaction score (average 1.6). Again the improvements were similar for both MSK and MH cases: 90% of MSK cases (N=3,166) improved by an average of 5.2, and 94% of MH cases (N=457) improved by an average of 5.0 (Figure 32). The mean score at entry changed from 2.9 to 7.5 at discharge for MSK cases, while from 2.5 to 7.2 for MH cases.

Figure 32: Change in COPM Satisfaction scores shown for MSK and MH cases

The percentage that improved was also similar when comparing those who were at work at entry (90%, N=2,829) with those who were absent (90%, N=925), see Figure 33. The size of the change of score was also similar for those who were absent at entry (an average of 5.6) and those who were at work (average of 5.0).

Figure 33: Change in COPM Satisfaction scores shown for those at work / absent at entry

4.6.8 Significant factors in the change in health measures

Multivariate analysis for change in health measures suggests that age is important and that younger cases had a greater improvement in their health measure scores ( EQ-5D index, HADS and COPM). Also, those cases with poorer health scores at entry, had a greater improvement in their health score.

The change in EQ-5D index score from entry to discharge is generally higher by 10% for men compared to women. Although both entry and discharge HADS scores are worse for MH cases than MSK cases, it is not surprising that MH cases improve their scores more during the intervention compared to the MSK cases. The change in COPM Satisfaction score is significantly different by Health Board area. The COPM satisfaction scores improved most in Dundee & Tayside, while the smallest changes were in Highland and Dumfries & Galloway. It is not clear whether this is due to improvements in cases' health or clinical variations.

4.7 Changes in medication

On discharge, cases reported whether they were still on the same medication they were taking for their primary health condition. For those who reported whether they were taking medication at both entry and discharge (N=4,942), 25% were not taking medication at either entry or discharge, while 22% were on the same medication as when they entered the programme. However, a third (33%) who had been taking medication at entry were not taking any medication at discharge, and 15% had reduced their medication use by the time they were discharged (Table 25).

Table 25: Changes in medication use from entry to discharge (N=4.942)

Entry, taking medication?		Discharge, taking medication?
		No	Same	Some reduction in meds	Some additional meds	Some additional / some reduction	Total
	No	21.6%	3.5%	3.2%	0.8%	0.4%	29.4%
	Yes	32.6%	18.4%	12.0%	4.9%	2.8%	70.6%
	Total	54.2%	21.8%	15.2%	5.7%	3.1%	100.0%

4.8 Use of other support services

Altogether 340 cases who were using additional support services at the time of entering the programme (e.g. medical professionals and allied health professionals) reported whether they were still using these when they were discharged from the service. Of these, over half (53%) were no longer using these services, 31% were using them the same amount, 15% were using them less and 2% were using them more.

4.9 Ability to work

Cases were asked at both entry and discharge whether they were working their normal hours, restricted hours or were off work. Of the 6,759 who answered the question at both entry and discharge, almost two thirds (64%) were working their normal hours at both times (Table 26), 19% who were off work at entry were working normal hours at discharge, while 5% who were on restricted hours at entry were working normal hours at discharge, meaning an improvement in working hours for almost a quarter of these cases.

Table 26: Hours worked at entry and discharge (N=6,759)

		Discharge
		Normal hours	Restricted hours	Off work	Total
Entry	Normal hours	64.2%	1.5%	1.2%	66.9%
	Restricted hours	5.1%	1.1%	0.3%	6.5%
	Off work	18.8%	2.4%	5.4%	26.7%
	Total	88.1%	5.0%	7.0%	100.0%

Cases were also asked whether they were able to do their normal duties at both entry and discharge. Only those who were at work at entry answered this question. Of the 4,940 who provided an answer at both entry and discharge 21% did not have difficulty with work duties at either entry or discharge (Table 27). However, 59% improved from struggling with their normal duties to doing their normal duties without difficulty. Furthermore, 4% also improved from not able to do their normal duties to being able to do them without difficult at discharge.

Table 27: Ability to perform work duties (N=4,940)

		Discharge
		Normal duties, no difficulty	Normal duties, but struggling	Not able to do normal duties	Total
Entry	Normal duties, no difficulty	20.8%	1.9%	0.2%	22.8%
	Normal duties, but struggling	58.7%	12.4%	0.6%	71.7%
	Not able to do normal duties	3.8%	1.4%	0.2%	5.5%
	Total	83.3%	15.7%	1.0%	100.0%

4.10 Prediction of ability to do job in 6 months' time

Cases were asked at entry and discharge whether, considering their health, they thought they would be able to do their job in 6 months' time. Altogether 5,969 provided an answer at both entry and discharge (Table 28). Two thirds of cases (66%) thought they would be able to do their job in 6 months' time, both at entry and discharge. A fifth (20%) changed from being unsure at entry (17%) or thinking they could not (3%) to thinking they could when they were discharged from the programme. There were 4% who at entry thought they could do their job in 6 months' time, but were unsure at discharge.

Table 28: Prediction of ability to do job in 6 months' time

Entry		Discharge
		Yes	Don't know	No	Total
	Yes	65.7%	4.0%	1.0%	70.7%
	Don't know	17.3%	7.1%	1.2%	25.6%
	No	2.5%	0.7%	0.5%	3.7%
	Total	85.5%	11.9%	2.6%	100.0%

4.11 Impact of the service after discharge

4.11.1 EQ-5D index

Altogether 2,033 cases provided EQ-5D data 3 months, 6 months, or both 3 and 6 months after discharge. The mean EQ-5D index scores were calculated for these cases at the different time points, and are shown in Figure 34, with those who completed the EQ-5D at all 4 points (blue), at entry, discharge and 3 months post discharge (green), and at entry, discharge and 6 months post discharge (yellow). The dates of completion of the follow-up questionnaires were not available, but it is assumed that they were approximately 3 and 6 months following discharge.

These figures show that the improvement in mean EQ-5D index score from entry to discharge was maintained at 3 and 6 months.

Figure 34: EQ-5D Index score for entry, discharge, 3 and 6 months post discharge

4.11.2 VAS score

A similar pattern was seen when considering the visual analogue scale scores of overall health (where a score of 100 represents the best health imaginable) as shown in Figure 35. The improvement in health appears to be largely maintained 3 and 6 months following discharge from the programme.

Figure 35: VAS scores at entry, discharge, 3 months and 6 months post discharge

4.11.3 Ability to work normal hours

Altogether, 1,959 cases answered a question about their ability to work their normal hours at entry, discharge and 3 and / or 6 months post-discharge. Where a response was received both at 3 and 6 months post-discharge, the 6 month response was taken, as a better indication of the durability of the impact of the service. Thus a case's ability to work normal hours is available at 3 time points (entry, discharge and post discharge), and can be classed as 'work' (working normal hours), 'restricted' (working restricted hours) and 'off' (off work). Figure 36 shows the proportion of respondents falling into the different categories of work ability at the three time points.

Almost two thirds of cases (65%) were working their normal hours at all three time points, while a further 23% who were off work or restricted at entry had returned to normal working hours at discharge and remained working normal hours at follow up.

Of the 1,794 who were working normal hours at discharge, 96.1% were still working normal hours at follow up; 3.1% were working restricted hours, while 0.8% were off sick at follow up. It appears that the vast majority have been able to maintain their normal working hours after leaving the programme.

Figure 36: Ability to work normal hours at entry, discharge and follow-up (N=1,959)

4.12 Qualitative feedback on the service

Discharged cases' views of the service are shown in Table 29. Over 98% of cases made a positive comment about all dimensions relating to the service delivery (questions 1-8). Over 99% reported it as a good or excellent experience and good or excellent in terms of its helpfulness; and 99% would recommend the service to others and would use it again.

The last three questions (9-11) related to the impact of the service on the cases' ability to remain in work or return to work. Altogether 93% agreed with the statement 'This programme has had a positive impact on my current work situation" (question 9).

When asked 'Do you think the service had helped you to stay in work or be closer to getting back to work?' 87% replied yes, while 8% were unsure (question 10). There were just 5% who answered no, implying that the service was seen as beneficial in helping to maintain the ability to work by the majority of cases.

A second, similar question was also asked (question 11): 'Do you feel the service helped you to return to work more quickly than if you had not had the support of the service?'. Unfortunately, neither a 'not applicable' nor 'don't know' option were not offered for this question. However, no answer was provided by more than 2,000 cases who had answered the rest of the feedback questions, implying that the question was not relevant for them. Of those who did answer this question, 605 cases provided a comment that they had not been off sick, so their responses were excluded from the analysis. Of the remaining cases who answered this question (3,042), 85% thought the service helped them to return to work more quickly than if they had not had the support of the service, which is a very similar proportion to the responses to question 10.

Table 29: Subjective feedback on service provision

Question	Positive response	%	N
1. How would you rate your overall experience of the service?	Good or excellent	99.4%	5,708
2. How helpful was the support you received?	Good or excellent	99.4%	5,700
3. How involved did you feel throughout the entire process?	Good or excellent	99.3%	5,649
4. How would you rate the treatment you have received?	Good or excellent	98.6%	5,597
5. How would you rate the venue you were seen in?	Good or excellent	98.3%	5,567
6. How would you rate the speed and delivery of this service?	Good or excellent	98.6%	5,665
7. I would recommend this service to others	Agree	99.0%	5,599
8. I would use this service again	Agree	98.7%	5,638
9. This programme has had a positive impact on my current work situation	Agree	93.4%	5,540
10. Do you think the service has helped you to stay in work or be closer to getting back to work?	Yes Don't know	86.5% 8.1%	5,303
11. Do you feel the service helped you to return to work more quickly than if you had not had the support of the service?	Yes	84.6%	3,042