Research and analysis

Health Assessment Channels Trial: technical annex

Published 15 September 2026

1. Background and methodology

This technical annex accompanies the Health Assessment Channels Trial (HACT) report and contains methodological detail supporting the published findings.

2. Trial Design and analytical framework

The HACT was conducted between May 2022 and January 2023. Eligible claimants were those undertaking an initial assessment, capable of using all 3 channels and suitable for auto-allocation. Results were assessed using both Intention to Treat (allocated channel) and Per Protocol (attended channel) analytical frameworks.

3. Data sources

Analysis combined assessment provider booking and referral data, the Department for Work and Pensions (DWP) administrative benefit outcome datasets, and customer research evidence. Outcomes include subsequent mandatory reconsideration and appeals, where applicable.

4. Limitations

Attended-channel analysis should be interpreted with caution because channel changing may introduce selection effects.

5. Statistical methodologies

Balance testing used Standardised Mean Differences (SMDs). Award outcomes were tested using chi-square and Fisher exact tests where appropriate. Award lengths and prognosis periods were assessed using independent t-tests. Logistic regression was used for channel change analysis.

Abbreviations

Abbreviation	Meaning
CI	Confidence interval
Df	Degrees of freedom
pp	Percentage points
EXP	Exponentiated Beta Coefficient
ICD10	International Classification of Diseases, Tenth Revision (ICD-10) condition
t	t-statistic

Section A – Standardised Mean Difference tests

SMD tests were used to assess whether claimants allocated to different assessment channels were broadly comparable in terms of observable characteristics prior to the start of the Health Assessment Channels Trial.

SMDs measure the difference in means or proportions between groups, expressed relative to the pooled standard deviation, and are commonly used to assess balance between groups in experimental and quasi‑experimental designs.

For the trial analysis, SMDs were calculated for key demographic, geographic and health‑related characteristics, including age, sex, region, recorded condition groups, and Index of Multiple Deprivation (IMD) domains.

As a rule of thumb, absolute SMD values between -0.1 and 0.1 are generally considered to indicate good balance between groups.

The results presented in this appendix show that, among trial‑eligible claimants who were automatically allocated to a channel, differences across channels were small in magnitude, supporting the assumption that the allocation process did not introduce material population bias and that outcomes can be meaningfully compared across assessment channels.

Figure A.1 IAS IMD scores: Claimants eligible for all channels and auto-allocated

Figure A.2 IAS characteristics: Eligible for all channels and auto-allocated

Figure A.3 IAS IMD scores: Eligible for all channels and auto or manually allocated

Figure A.4 IAS characteristics: Eligible for all channels and auto or manually allocated

Figure A.5 Capita IMD scores: Eligible for all channels and auto-allocated

Figure A.6 Capita characteristics: Eligible for all channels and auto-allocated

Figure A.7 Capita IMD scores: Eligible for all channels and auto or manually allocated

Figure A.8 Capita characteristics: Eligible for all channels and auto or manually allocated

Figure A.9 Centre for Health and Disability Assessments (CHDA) IMD scores: Eligible for all channels and auto-allocated

Figure A.10 CHDA characteristics: Eligible for all channels and auto-allocated

Figure A.11 CHDA IMD scores: Eligible for all channels and auto or manually allocated

Figure A.12 CHDA characteristics: Eligible for all channels and auto or manually allocated

Section B – Personal Independence Payment (PIP) allocated channel analyses

Annex B presents Chi-square tests of association underpinning the analysis of award rates by allocated channel, stratified by recorded primary condition group, gender and age.

Award rates are shown for each channel within each category. Statistical significance is assessed at the 5% level (p ≤ 0.05) and effect sizes are provided using Cramer’s V. Where expected cell counts are small, Chi-square results may be unreliable, and Fisher’s exact test should be used.

Independent t-tests have been used to test association between channel and award length. Statistical significance is assessed at the 5% level (p ≤ 0.05).

Assessments from Scotland and Northern Ireland were excluded from the results due to devolution of powers to the local government, in these regions.

Table B.1 PIP outcomes by allocated channel

Allocated channel	Awarded	Disallowed	Award rate per cent
F2F	22,231	22,123	50.1
VA	19,207	18,877	50.4
TA	84,178	86,810	49.2

Table B.2 Tests of association between PIP award and allocated channel

Comparison	Test	Statistic	Df	Probability-value
F2F vs VA	Pearson chi‑square	0.7953	1	0.373
F2F vs VA	Fisher’s exact (two‑sided)	–	–	0.373
F2F vs VA	Cramer’s V	0.0031	–	–
F2F vs TA¹	Pearson chi‑square	11.1951	1	0.001
F2F vs TA¹	Fisher’s exact (two‑sided)	–	–	0.002
F2F vs TA	Cramer’s V	0.0072	–	–

¹ Result is statistically significant at (or below) the .05 level

Table B.3 PIP outcomes by allocated channel (F2F vs TA) and by primary condition

Primary condition group	F2F awarded	F2F disallowed	TA awarded	TA disallowed
Anxiety and depression	4,806	4,546	23,311	22,446
Psychiatric disorders excluding anxiety and depression	721	586	2,097	1,671
Musculoskeletal disease (general)	4,645	3,519	16,870	13,020
Musculoskeletal disease (regional)	3,286	3,047	13,282	11,622
Neurological disease	2,709	2,116	8,256	6,647
Respiratory disease	1,004	1,490	3,160	5,000
Cardiovascular disease	808	968	2,842	3,505
Malignant disease	680	626	1,270	1,163
Autism	595	684	1,179	1,148
Attention deficit hyperactivity disorder (ADHD) or attention deficit disorder (ADD)	373	660	1,211	2,456
Endocrine disease	340	752	1,461	3,044
Gastrointestinal disease	330	820	1,345	3,657
Learning disability	132	122	220	266
Unknown	2	7	974	2,980
All other categories	1,800	2,180	6,700	8,185

Table B.4 Test of association between PIP award and allocated channel (F2F vs TA) by primary condition

Primary condition group	F2F award rate per cent	TA award rate per cent	Difference (pp)	Probability-value	Cramer’s V
Musculoskeletal disease (general)	56.9	56.4	-0.5	0.4615	0.0038
Musculoskeletal disease (regional)¹	51.9	53.3	1.4	0.0395	0.0116
Neurological disease	56.1	55.4	-0.7	0.3642	0.0065
Malignant disease	52.1	52.2	0.1	0.9388	0.0013
Learning disability	52.0	45.3	-6.7	0.0831	0.0637
Anxiety and depression	51.4	50.9	-0.5	0.4329	0.0033
Psychiatric disorders excluding anxiety and depression	55.2	55.7	0.5	0.7595	0.0043
Autism¹	46.5	50.7	4.2	0.0172	0.0397
Cardiovascular disease	45.5	44.8	-0.7	0.5906	0.0060
Respiratory disease	40.3	38.7	-1.6	0.1702	0.0133
ADHD or ADD	36.1	33.0	-3.1	0.0640	0.0270
Endocrine disease	31.1	32.4	1.3	0.4111	0.0110
Gastrointestinal disease	28.7	26.9	-1.8	0.2146	0.0158
Unknown†	22.2	24.6	2.4	0.8668	0.0027
All other categories	45.2	45.0	-0.2	0.8092	0.0018

¹ Result is statistically significant at (or below) the .05 level.

† Where expected cell counts are small, Chi-square assumptions may not hold; Fisher’s exact test should be used and results interpreted cautiously.

Table B.5 PIP outcomes by allocated channel (F2F vs VA) and by primary condition

Primary condition group	F2F awarded	F2F disallowed	VA awarded	VA disallowed
Anxiety and depression	4,806	4,546	5,232	4,893
Psychiatric disorders excluding anxiety and depression	721	586	523	394
Musculoskeletal disease (general)	4,645	3,519	3,775	2,767
Musculoskeletal disease (regional)	3,286	3,047	2,899	2,434
Neurological disease	2,709	2,116	2,026	1,558
Respiratory disease	1,004	1,490	705	1,039
Cardiovascular disease	808	968	659	743
Malignant disease	680	626	297	273
Autism	595	684	289	351
ADHD or ADD	373	660	307	582
Endocrine disease	340	752	395	670
Gastrointestinal disease	330	820	323	764
Learning disability	132	122	76	64
Unknown	2	7	150	547
All other categories	1,800	2,180	1,551	1,798

Table B.6 Test of association between PIP award and allocated channel (F2F vs VA) by primary condition

Primary condition group	F2F award rate per cent	VA award rate per cent	Difference (pp)	Probability-value	Cramer’s V
Musculoskeletal disease (general)	56.9	57.7	0.8	0.3250	0.0081
Musculoskeletal disease (regional)¹	51.9	54.4	2.5	0.0077	0.0247
Neurological disease	56.1	56.5	0.4	0.7256	0.0038
Anxiety and depression	51.4	51.7	0.3	0.6919	0.0028
Psychiatric disorders excluding anxiety and depression	55.2	57.0	1.8	0.3821	0.0185
Malignant disease	52.1	52.1	0.0	0.9879	0.0003
Learning disability	52.0	54.3	2.3	0.6592	0.0222
Autism	46.5	45.2	-1.3	0.5718	0.0129
Cardiovascular disease	45.5	47.0	1.5	0.3969	0.0150
Respiratory disease	40.3	40.4	0.1	0.9128	0.0017
ADHD or ADD	36.1	34.5	-1.6	0.4714	0.0164
Endocrine disease¹	31.1	37.1	6.0	0.0035	0.0628
Gastrointestinal disease	28.7	29.7	1.0	0.5962	0.0112
Unknown†	22.2	21.5	-0.7	0.9594	0.0019
All other categories	45.2	46.3	1.1	0.3524	0.0109

¹ Result is statistically significant at (or below) the .05 level.

† Where expected cell counts are small, Chi-square assumptions may not hold; Fisher’s exact test should be used and results interpreted cautiously.

Table B.7 PIP outcomes by allocated channel and by gender

Gender	Allocated channel	Awarded	Disallowed	Awarded per cent
Female	F2F	13,131	12,842	50.6
Female	VA	11,653	11,187	51.0
Female	TA	50,727	51,042	49.8
Male	F2F	9,100	9,281	49.5
Male	VA	7,554	7,690	49.6
Male	TA	33,449	35,768	48.3

Table B.8 Test of association between PIP award and allocated channel by gender

Comparison	Gender	F2F award rate per cent	Comparator award rate per cent	Difference (pp)	Probability-value	Cramer’s V
F2F vs VA	Female	50.6	51.0	−0.46	0.3065	0.0046
F2F vs VA	Male	49.5	49.6	−0.04	0.9327	0.0005
F2F vs TA¹	Female	50.6	49.8	0.71	0.0408	0.0057
F2F vs TA¹	Male	49.5	48.3	1.19	0.0043	0.0096

¹ Result is statistically significant at (or below) the .05 level.

Table B.9 PIP outcomes by allocated channel and by age

Age band	Channel	Awarded	Disallowed	Awarded per cent
16 to 19	F2F	750	847	47.0
16 to 19	VA	578	695	45.4
16 to 19	TA	2,514	2,917	46.3
20 to 24	F2F	1,166	1,661	41.2
20 to 24	VA	1,016	1,499	40.4
20 to 24	TA	4,635	6,665	41.0
25 to 29	F2F	1,371	1,817	43.0
25 to 29	VA	1,375	1,849	42.6
25 to 29	TA	5,905	8,166	42.0
30 to 34	F2F	1,851	2,241	45.2
30 to 34	VA	1,758	2,026	46.5
30 to 34	TA	7,638	9,545	44.5
35 to 39	F2F	1,952	2,066	48.6
35 to 39	VA	1,857	2,023	47.9
35 to 39	TA	8,225	9,183	47.2
40 to 44	F2F	1,980	2,155	47.9
40 to 44	VA	1,944	1,874	50.9
40 to 44	TA	8,415	8,808	48.9
45 to 49	F2F	2,233	2,082	51.7
45 to 49	VA	1,902	1,812	51.2
45 to 49	TA	8,525	8,096	51.3
50 to 54	F2F	2,925	2,644	52.5
50 to 54	VA	2,471	2,108	54.0
50 to 54	TA	10,683	10,018	51.6
55 to 59	F2F	3,405	2,977	53.4
55 to 59	VA	2,794	2,274	55.1
55 to 59	TA	12,180	10,564	53.6
60 to 64	F2F	3,808	3,023	55.7
60 to 64	VA	2,916	2,251	56.4
60 to 64	TA	12,729	10,690	54.4
65 to 69	F2F	787	610	56.3
65 to 69	VA	596	466	56.1
65 to 69	TA	2,725	2,156	55.8

Table B.10 Test of association between PIP award and allocated channel by age

Comparison	Age band	F2F awarded per cent	Comparator awarded per cent	Difference (pp)	Probability-value	Cramer’s V
F2F vs VA	16 to 19	47.0	45.4	1.56	0.4055	0.016
F2F vs VA	20 to 24	41.3	40.4	0.85	0.5293	0.009
F2F vs VA	25 to 29	43.0	42.7	0.36	0.7732	0.004
F2F vs VA	30 to 34	45.2	46.5	−1.23	0.2760	0.012
F2F vs VA	35 to 39	48.6	47.9	0.72	0.5217	0.007
F2F vs VA¹	40 to 44	47.9	50.9	−3.04	0.0069	0.030
F2F vs VA	45 to 49	51.8	51.2	0.54	0.6305	0.005
F2F vs VA	50 to 54	52.5	54.0	−1.44	0.1478	0.014
F2F vs VA	55 to 59	53.4	55.1	−1.78	0.0580	0.018
F2F vs VA	60 to 64	55.8	56.4	−0.69	0.4514	0.007
F2F vs VA	65 to 69	56.3	56.1	0.22	0.9154	0.002
F2F vs TA	16 to 19	47.0	46.3	0.67	0.6353	0.006
F2F vs TA	20 to 24	41.3	41.0	0.23	0.8260	0.002
F2F vs TA	25 to 29	43.0	42.0	1.04	0.2833	0.008
F2F vs TA	30 to 34	45.2	44.5	0.78	0.3647	0.006
F2F vs TA	35 to 39	48.6	47.3	1.33	0.1272	0.010
F2F vs TA	40 to 44	47.9	48.9	−0.98	0.2599	0.008
F2F vs TA	45 to 49	51.8	51.3	0.46	0.5908	0.004
F2F vs TA	50 to 54	52.5	51.6	0.91	0.2242	0.008
F2F vs TA	55 to 59	53.4	53.6	−0.20	0.7778	0.002
F2F vs TA¹	60 to 64	55.8	54.4	1.4	0.0419	0.012
F2F vs TA	65 to 69	56.3	55.8	0.51	0.7368	0.004

¹ Result is statistically significant at (or below) the .05 level.

Table B.11 PIP award length by allocated channel

Allocated channel	N	Mean (months)	Standard deviation
F2F	17,604	34.48	8.58
VA	15,623	33.56	8.09
TA	68,973	33.43	7.83

Table B.12 Test of association between PIP award length and allocated channel

Comparison	Test	Mean diff (months)	95% CI	t	Probability-value
F2F vs VA¹	Satterthwaite	0.92	0.75 – 1.10	10.11	<0.0001
F2F vs TA¹	Satterthwaite	1.05	0.91 – 1.19	14.76	<0.0001

¹ Result is statistically significant at (or below) the .05 level.

Section C – PIP attended channel analyses

Annex C presents Chi-square tests of association underpinning the analysis of award rates by attended channel, stratified by recorded primary condition group, gender and age.

Award rates are shown for each channel within each category. Statistical significance is assessed at the 5% level (p ≤ 0.05) and effect sizes are provided using Cramer’s V. Where expected cell counts are small, Chi-square results may be unreliable, and Fisher’s exact test should be used.

Independent t-tests have been used to test for association between channel and award length (PIP). Statistical significance is assessed at the 5% level (p ≤ 0.05).

Assessments from Scotland and Northern Ireland were excluded from the results due to devolution of powers to the local government, in these regions.

Table C.1 PIP outcomes by attended channel

Attended channel	Awarded	Disallowed	Award rate per cent
F2F	15,301	16,440	48.2
VA	14,378	14,551	49.7
TA	95,936	96,819	49.8

Table C.2 Test of association between PIP award and attended channel

Comparison	Test	Statistic	Df	Probability-value
F2F vs VA¹	Pearson chi‑square	13.5407	1	0.0002
F2F vs VA¹	Fisher’s exact (two‑sided)	–	–	0.0002
F2F vs VA	Cramer’s V	0.0149	–	–
F2F vs TA¹	Pearson chi‑square	26.7074	1	<0.0001
F2F vs TA¹	Fisher’s exact (two‑sided)	–	–	<0.0001
F2F vs TA	Cramer’s V	0.0109	–	–

¹ Result is statistically significant at (or below) the .05 level.

Table C.3 PIP outcomes by attended channel (F2F vs TA) and by primary condition

Primary condition group	F2F awarded	F2F disallowed	TA awarded	TA disallowed
Musculoskeletal disease (general)	3,337	2,735	19,111	14,457
Musculoskeletal disease (regional)	2,370	2,376	14,883	12,857
Anxiety and depression	2,950	3,023	26,613	25,100
Psychiatric disorders excluding anxiety and depression	508	439	2,489	1,941
Neurological disease	1,909	1,594	9,497	7,491
Respiratory disease	698	1,153	3,671	5,563
Cardiovascular disease	543	763	3,271	3,920
Malignant disease	461	453	1,567	1,392
Autism	438	508	1,381	1,374
Endocrine disease	252	591	1,668	3,379
ADHD or ADD	250	458	1,396	2,756
Gastrointestinal disease	217	622	1,532	4,018
Learning disability	103	91	276	317
Unknown	3	9	1006	3,131
All other categories	1,262	1,625	7,575	9,123

Table C.4 Test of association between PIP award and attended channel (F2F vs TA) by primary condition

Primary condition group	F2F award rate per cent	TA award rate per cent	Difference (pp)	Probability-value	Cramer’s V
Musculoskeletal disease (general)¹	55.0	56.9	2.0	0.004	0.0140
Musculoskeletal disease (regional)¹	49.9	53.7	3.8	<0.001	0.0260
Neurological disease	54.5	55.9	1.4	0.127	0.0110
Anxiety and depression	49.4	51.5	2.1	0.002	0.0130
Psychiatric disorders excluding anxiety and depression	53.6	56.2	2.6	0.153	0.0200
Learning disability	53.1	46.5	−6.6	0.113	0.0570
Malignant disease	50.4	53.0	2.5	0.183	0.0210
Autism¹	46.3	50.1	3.8	0.042	0.0330
Cardiovascular disease¹	41.6	45.5	3.9	0.009	0.0280
Respiratory disease	37.7	39.8	2.0	0.100	0.0160
ADHD or ADD	35.3	33.6	−1.7	0.380	0.0130
Endocrine disease	29.9	33.1	3.2	0.070	0.0240
Gastrointestinal disease	25.9	27.6	1.7	0.292	0.0130
Unknown†	25.0	24.3	−0.7	0.956	0.0010
All other categories	43.7	45.4	1.6	0.100	0.0120

¹ Result is statistically significant at (or below) the .05 level.

† Where expected cell counts are small, Chi-square assumptions may not hold; Fisher’s exact test should be used and results interpreted cautiously.

Table C.5 PIP outcomes by attended channel (F2F vs VA) and by primary condition

Primary condition group	F2F awarded	F2F disallowed	VA awarded	VA disallowed
Musculoskeletal disease (general)	3,337	2,735	2,842	2,114
Musculoskeletal disease (regional)	2,370	2,376	2,214	1,870
Anxiety and depression	2,950	3,023	3,786	3,762
Psychiatric disorders excluding anxiety and depression	508	439	344	271
Neurological disease	1,909	1,594	1,585	1,236
Respiratory disease	698	1,153	500	813
Cardiovascular disease	543	763	495	533
Malignant disease	461	453	219	217
Autism	438	508	244	301
Endocrine disease	252	591	276	496
ADHD or ADD	250	458	245	484
Gastrointestinal disease	217	622	249	601
Learning disability	103	91	49	44
Unknown	3	9	117	394
All other categories	1,262	1,625	1,213	1,415

Table C.6 Test of association between PIP award and attended channel (F2F vs VA) by primary condition

Primary condition group	F2F award rate per cent	VA award rate per cent	Difference (pp)	Probability-value	Cramer’s V
Musculoskeletal disease (general)¹	55.0	57.3	2.3	0.012	0.024
Musculoskeletal disease (regional)¹	49.9	54.2	4.3	<0.001	0.043
Neurological disease	54.5	56.2	1.7	0.179	0.017
Anxiety and depression	49.4	50.2	0.8	0.374	0.008
Psychiatric disorders excluding anxiety and depression	53.6	55.9	2.3	0.374	0.023
Learning disability	53.1	52.7	−0.4	0.949	0.004
Malignant disease	50.4	50.2	−0.2	0.943	0.002
Autism	46.3	44.8	−1.5	0.568	0.015
Cardiovascular disease¹	41.6	48.2	6.6	0.002	0.066
Respiratory disease	37.7	38.1	0.4	0.832	0.004
ADHD or ADD	35.3	33.6	−1.7	0.497	0.018
Endocrine disease¹	29.9	35.8	5.9	0.012	0.062
Gastrointestinal disease	25.9	29.3	3.4	0.115	0.038
Unknown†	25.0	22.9	−2.1	0.864	0.008
All other categories	43.7	46.2	2.5	0.068	0.025

¹ Result is statistically significant at (or below) the .05 level.

† Where expected cell counts are small, Chi-square assumptions may not hold; Fisher’s exact test should be used and results interpreted cautiously.

Table C.7 PIP outcomes by attended channel and by gender

Gender	Attended channel	Awarded	Disallowed	Awarded per cent
Female	F2F	8,782	9,251	48.7
Female	VA	8,912	8,772	50.4
Female	TA	57,816	57,048	50.3
Male	F2F	6,519	7,189	47.6
Male	VA	5,466	5,779	48.6
Male	TA	38,118	39,771	48.9

Table C.8 Test of association between PIP award and attended channel by gender

Comparison	Gender	F2F award rate per cent	Comparator award rate per cent	Difference (pp)	Probability-value	Cramer’s V
F2F vs VA¹	Female	48.7	50.4	−1.70	0.0013	0.0170
F2F vs VA	Male	47.6	48.6	−1.05	0.0979	0.0105
F2F vs TA¹	Female	48.7	50.3	−1.63	<0.0001	0.0112
F2F vs TA¹	Male	47.6	48.9	−1.38	0.0028	0.0099

¹ Result is statistically significant at (or below) the .05 level.

Table C.9 PIP outcomes by attended channel and by age

Age band	Attended Channel	Awarded	Disallowed	Awarded per cent
16 to 19	F2F	533	639	45.5
16 to 19	VA	458	594	43.5
16 to 19	TA	2,851	3,226	46.9
20 to 24	F2F	783	1,173	40.0
20 to 24	VA	829	1,260	39.7
20 to 24	TA	5,205	7,392	41.3
25 to 29	F2F	875	1,225	41.7
25 to 29	VA	1,099	1,516	42.0
25 to 29	TA	6,677	9,091	42.4
30 to 34	F2F	1,166	1,566	42.7
30 to 34	VA	1,403	1,702	45.2
30 to 34	TA	8,677	10,544	45.1
35 to 39	F2F	1,238	1,455	46.0
35 to 39	VA	1,415	1,651	46.2
35 to 39	TA	9,381	10,166	48.0
40 to 44	F2F	1,288	1,562	45.2
40 to 44	VA	1,473	1,455	50.3
40 to 44	TA	9,578	9,820	49.4
45 to 49	F2F	1,532	1,595	49.0
45 to 49	VA	1,454	1,401	50.9
45 to 49	TA	9,674	8,994	51.8
50 to 54	F2F	1,998	2,027	49.6
50 to 54	VA	1,796	1,576	53.3
50 to 54	TA	12,285	11,167	52.4
55 to 59	F2F	2,461	2,296	51.7
55 to 59	VA	2,049	1,611	56.0
55 to 59	TA	13,869	11,908	53.8
60 to 64	F2F	2,837	2,410	54.1
60 to 64	VA	1,996	1,490	57.3
60 to 64	TA	14,620	12,064	54.8
65 to 69	F2F	588	492	54.4
65 to 69	VA	406	295	57.9
65 to 69	TA	3,114	2,445	56.0

Table C.10 Test of association between PIP award and attended channel by age

Comparison	Age band	F2F awarded per cent	Comparator awarded per cent	Difference (pp)	Probability-value	Cramer’s V
F2F vs VA	16 to 19	45.5	43.5	−2.0	0.3580	0.020
F2F vs VA	20 to 24	40.0	39.7	−0.3	0.8220	0.004
F2F vs VA	25 to 29	41.7	42.0	0.4	0.8030	0.004
F2F vs VA	30 to 34	42.7	45.2	2.5	0.0540	0.025
F2F vs VA	35 to 39	46.0	46.2	0.2	0.8910	0.002
F2F vs VA¹	40 to 44	45.2	50.3	5.1	<0.0001	0.051
F2F vs VA	45 to 49	49.0	50.9	1.9	0.1350	0.019
F2F vs VA¹	50 to 54	49.6	53.3	3.6	0.0020	0.036
F2F vs VA¹	55 to 59	51.7	56.0	4.3	0.0001	0.042
F2F vs VA¹	60 to 64	54.1	57.3	3.2	0.0030	0.031
F2F vs VA	65 to 69	54.4	57.9	3.5	0.1490	0.034
F2F vs TA	16 to 19	45.5	46.9	1.4	0.3670	0.011
F2F vs TA	20 to 24	40.0	41.3	1.3	0.2810	0.009
F2F vs TA	25 to 29	41.7	42.4	0.7	0.5540	0.004
F2F vs TA¹	30 to 34	42.7	45.1	2.4	0.0150	0.016
F2F vs TA¹	35 to 39	46.0	48.0	2.0	0.0490	0.013
F2F vs TA¹	40 to 44	45.2	49.4	4.2	<0.0001	0.028
F2F vs TA¹	45 to 49	49.0	51.8	2.8	0.0030	0.020
F2F vs TA¹	50 to 54	49.6	52.4	2.7	0.0010	0.019
F2F vs TA¹	55 to 59	51.7	53.8	2.1	0.0090	0.015
F2F vs TA	60 to 64	54.1	54.8	0.7	0.3380	0.005
F2F vs TA	65 to 69	54.4	56.0	1.6	0.3410	0.012

¹ Result is statistically significant at (or below) the .05 level.

Table C.11 PIP award length by attended channel

Attended channel	N	Mean (months)	Standard deviation
F2F	11,915	34.70	8.81
VA	11,772	33.53	8.17
TA	78,512	33.48	7.85

Table C.12 Test of association between PIP award length and attended channel

Comparison	Test	Mean diff (months)	95% CI	t	Probability-value
F2F vs VA¹	Satterthwaite	1.18	0.96 – 1.39	10.67	<0.0001
F2F vs TA¹	Satterthwaite	1.22	1.05 – 1.39	14.30	<0.0001

¹ Result is statistically significant at (or below) the .05 level.

Section D – Work Capability Assessment (WCA) allocated channel analyses

Annex D presents Chi-square tests of association underpinning the analysis of health award rates by allocated channel, stratified by first recorded ICD10 medical condition, gender and age.

Decision rates are shown for each channel within condition. Statistical significance is assessed at the 5% level (p ≤ 0.05) and effect sizes are provided using Cramer’s V. Where expected cell counts are small, Chi-square results may be unreliable, and Fisher’s exact test should be used.

Independent t-tests have been used to test for association between allocated channel and prognosis period (WCA). Statistical significance is assessed at the 5% level (p ≤ 0.05).

Assessments from Scotland and Northern Ireland were excluded from the results to ensure consistency between the WCA and PIP.

Table D.1 WCA health award by allocated channel

Allocated channel	Health award	No health award	Health award per cent
F2F	16,061	4,814	76.9
VA	5,777	1,556	78.8
TA	54,805	15,379	78.1

Table D.2 Test of association between health award and allocated channel

Comparison	Test	Statistic	Df	Probability-value
F2F vs VA¹	Pearson chi‑square	10.5311	1	0.0012
F2F vs VA¹	Fisher’s exact (two‑sided)	–	–	0.0012
F2F vs VA	Cramer’s V	0.0193	–	–
F2F vs TA¹	Pearson chi‑square	12.3011	1	0.0005
F2F vs TA¹	Fisher’s exact (two‑sided)	–	–	0.0005
F2F vs TA	Cramer’s V	0.0116	–	–

¹ Result is statistically significant at (or below) the .05 level.

Table D.3 WCA health award by allocated channel (F2F vs TA) and first recorded ICD10 condition

First recorded ICD10 condition	Health award (F2F)	No health award (F2F)	Health award (TA)	No health award (TA)
Certain infectious and parasitic diseases	29	13	78	30
Congenital malformations, deformations and chromosomal abnormalities	14	2	40	1
Diseases of the blood and blood forming organs and certain diseases involving the immune mechanism	60	14	133	62
Diseases of the circulatory system	974	294	2,709	898
Diseases of the digestive system	459	217	1,606	667
Diseases of the ear and mastoid process	56	25	179	78
Diseases of the eye and adnexa	160	52	407	151
Diseases of the genitourinary system	253	85	713	238
Diseases of the musculoskeletal system and connective tissue	4,234	2,240	13,316	6,429
Diseases of the nervous system	674	176	2,172	534
Diseases of the respiratory system	733	155	2,515	457
Diseases of the skin and subcutaneous system	77	33	281	122
Endocrine, nutritional and metabolic diseases	361	168	1,082	513
Factors influencing health status and contact with health services	48	34	196	203
Injury, poisoning and certain other consequences of external causes	64	25	178	84
Mental and behavioural disorders	7,524	1,142	28,049	4,469
Neoplasms	171	40	502	111
Pregnancy, childbirth and the puerperium	0	0	5	4
Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified	123	57	445	155
Unknown or claimants without diagnosis on the system	0	0	1	0

Table D.4 Test of association between WCA health award and allocated channel (F2F vs TA) and first recorded ICD10 condition

First recorded ICD10 condition	F2F health award per cent	TA health award per cent	Difference (pp)	Probability-value	Cramer’s V
Congenital malformations, deformations and chromosomal abnormalities†	87.5	97.6	-10.06	0.126	0.2025
Mental and behavioural disorders	86.8	86.3	0.56	0.173	0.0067
Diseases of the respiratory system	82.6	84.6	-2.07	0.137	0.0239
Diseases of the blood and blood forming organs and certain diseases involving the immune mechanism¹	81.1	68.2	12.87	0.036	0.1277
Neoplasms	81.0	81.9	-0.85	0.783	0.0096
Diseases of the nervous system	79.3	80.3	-0.98	0.536	0.0104
Diseases of the circulatory system	76.8	75.1	1.71	0.223	0.0175
Diseases of the eye and adnexa	75.5	72.9	2.53	0.476	0.0257
Diseases of the genitourinary system	74.9	75.0	-0.12	0.965	0.0012
Injury, poisoning and certain other consequences of external causes	71.9	67.9	3.97	0.484	0.0373
Diseases of the skin and subcutaneous system	70.0	69.7	0.27	0.956	0.0024
Diseases of the ear and mastoid process	69.1	69.7	-0.51	0.930	0.0048
Certain infectious and parasitic diseases	69.1	72.2	-3.17	0.700	0.0315
Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified	68.3	74.2	-5.84	0.123	0.0552
Endocrine, nutritional and metabolic diseases	68.2	67.8	0.40	0.863	0.0038
Diseases of the digestive system	67.9	70.7	-2.76	0.170	0.0253
Diseases of the musculoskeletal system and connective tissue¹	65.4	67.4	-2.04	0.003	0.0187
Factors influencing health status and contact with health services	58.5	49.1	9.42	0.120	0.0708

¹ Result is statistically significant at (or below) the .05 level.

† Where expected cell counts are small, Chi-square assumptions may not hold; Fisher’s exact test should be used and results interpreted cautiously.

Tests for ‘Pregnancy, Childbirth and the puerperium’ and ‘Unknown or claimants without diagnosis on the system’ were not undertaken since there was no comparator to test.

Table D.5 WCA health award by allocated channel (F2F vs VA) and first recorded ICD10 condition

First recorded ICD10 condition	Health award (F2F)	No health award (F2F)	Health award (VA)	No health award (VA)
Certain infectious and parasitic diseases	29	13	8	2
Congenital malformations, deformations and chromosomal abnormalities	14	2	4	1
Diseases of the blood and blood forming organs and certain diseases involving the immune mechanism	60	14	16	5
Diseases of the circulatory system	974	294	296	97
Diseases of the digestive system	459	217	156	65
Diseases of the ear and mastoid process	56	25	14	8
Diseases of the eye and adnexa	160	52	55	15
Diseases of the genitourinary system	253	85	95	21
Diseases of the musculoskeletal system and connective tissue	4,234	2,240	1,402	624
Diseases of the nervous system	674	176	284	60
Diseases of the respiratory system	733	155	250	61
Diseases of the skin and subcutaneous system	77	33	36	12
Endocrine, nutritional and metabolic diseases	361	168	127	42
Factors influencing health status and contact with health services	48	34	27	16
Injury, poisoning and certain other consequences of external causes	64	25	18	7
Mental and behavioural disorders	7,524	1,142	2,843	468
Neoplasms	171	40	52	13
Pregnancy, childbirth and the puerperium	0	0	0	1
Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified	123	57	52	15
Unknown or claimants without diagnosis on the system	0	0	0	0

Table D.6 Test of association between WCA health award and allocated channel (F2F vs VA) and first recorded ICD10 condition

First recorded ICD10 condition	F2F health award per cent	VA health award per cent	Difference (pp)	Probability-value	Cramer’s V
Certain infectious and parasitic diseases†	69.1	80.0	-10.95	0.492	0.0953
Congenital malformations, deformations and chromosomal abnormalities†	87.5	80.0	7.50	0.676	0.0913
Diseases of the blood and blood forming organs and certain diseases involving the immune mechanism†	81.1	76.2	4.89	0.621	0.0507
Diseases of the circulatory system	76.8	75.3	1.49	0.541	0.0150
Diseases of the digestive system	67.9	70.6	-2.69	0.455	0.0250
Diseases of the ear and mastoid process	69.1	63.6	5.50	0.624	0.0483
Diseases of the eye and adnexa	75.5	78.6	-3.10	0.597	0.0315
Diseases of the genitourinary system	74.9	81.9	-7.05	0.122	0.0726
Diseases of the musculoskeletal system and connective tissue¹	65.4	69.2	-3.80	0.002	0.0343
Diseases of the nervous system	79.3	82.6	-3.27	0.200	0.0371
Diseases of the respiratory system	82.6	80.4	2.16	0.394	0.0246
Diseases of the skin and subcutaneous system	70.0	75.0	-5.00	0.522	0.0509
Endocrine, nutritional and metabolic diseases	68.2	75.2	-6.91	0.088	0.0645
Factors influencing health status and contact with health services	58.5	62.8	-4.25	0.645	0.0413
Injury, poisoning and certain other consequences of external causes	71.9	72.0	-0.09	0.993	0.0008
Mental and behavioural disorders	86.8	85.9	0.95	0.170	0.0125
Neoplasms	81.0	80.0	1.04	0.852	0.0112
Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified	68.3	77.6	-9.28	0.154	0.0908

¹ Result is statistically significant at (or below) the .05 level.

† Where expected cell counts are small, Chi-square assumptions may not hold; Fisher’s exact test should be used and results interpreted cautiously.

Tests for ‘Pregnancy, Childbirth and the puerperium’ and ‘Unknown or claimants without diagnosis on the system’ were not undertaken since there was no comparator to test.

Table D.7 WCA health award by allocated channel and by gender

Gender	Allocated channel	Health award	No health award	Health award per cent
Female	F2F	8,254	2,346	77.9
Female	VA	3,146	762	80.5
Female	TA	29,577	7,593	79.6
Male	F2F	7,807	2,468	76.0
Male	VA	2,631	794	76.8
Male	TA	25,228	7,786	76.4

Table D.8 Test of association between WCA health award and allocated channel by gender

Comparison	Gender	F2F health award per cent	Comparator health award per cent	Difference (pp)	Probability-value	Cramer’s V
F2F vs VA¹	Female	77.9	80.5	−2.63	0.0006	0.0285
F2F vs VA	Male	76.0	76.8	−0.84	0.3193	0.0085
F2F vs TA¹	Female	77.9	79.6	−1.70	0.0001	0.0174
F2F vs TA	Male	76.0	76.4	−0.44	0.3645	0.0044

¹ Result is statistically significant at (or below) the .05 level.

Table D.9 WCA health award by allocated channel and by age

Age band	Allocated channel	Health award	No health award	Health award per cent
Under 18	F2F	45	7	86.5
Under 18	VA	18	3	85.7
Under 18	TA	152	19	88.9
18 to 24	F2F	1,900	251	88.3
18 to 24	VA	700	103	87.2
18 to 24	TA	6,706	1,016	86.8
25 to 34	F2F	2,702	692	79.6
25 to 34	VA	1,206	267	81.9
25 to 34	TA	10,757	2,606	80.5
35 to 44	F2F	3,075	1,076	74.1
35 to 44	VA	1,189	354	77.1
35 to 44	TA	11,155	3,377	76.8
45 to 49	F2F	1,550	519	74.9
45 to 49	VA	597	167	78.1
45 to 49	TA	5,293	1,585	77.0
50 to 54	F2F	1,955	620	75.9
50 to 54	VA	623	191	76.5
50 to 54	TA	6,149	1,821	77.2
55 to 59	F2F	2,167	686	76.0
55 to 59	VA	713	209	77.3
55 to 59	TA	6,626	2,065	76.2
60 to 64	F2F	2,275	774	74.6
60 to 64	VA	626	203	75.5
60 to 64	TA	6,796	2,264	75.0
65 and over	F2F	392	189	67.5
65 and over	VA	105	59	64.0
65 and over	TA	1,171	626	65.2

Table D.10 Test of association between WCA health award and allocated channel by age

Comparison	Age band	F2F health award per cent	Comparator health award per cent	Difference (pp)	Probability-value	Cramer’s V
F2F vs VA†	Under 18	86.5	85.7	0.83	1.0000	0.0109
F2F vs VA	18 to 24	88.3	87.2	1.16	0.3886	0.0159
F2F vs VA	25 to 34	79.6	81.9	−2.26	0.0683	0.0261
F2F vs VA¹	35 to 44	74.1	77.1	−2.98	0.0212	0.0305
F2F vs VA	45 to 49	74.9	78.1	−3.22	0.0753	0.0334
F2F vs VA	50 to 54	75.9	76.5	−0.62	0.7207	0.0061
F2F vs VA	55 to 59	76.0	77.3	−1.37	0.3928	0.0139
F2F vs VA	60 to 64	74.6	75.5	−0.90	0.5974	0.0085
F2F vs VA	65 and over	67.5	64.0	3.45	0.4083	0.0303
F2F vs TA	Under 18	86.5	88.9	−2.35	0.6437	0.0310
F2F vs TA	18 to 24	88.3	86.8	1.49	0.0680	0.0184
F2F vs TA	25 to 34	79.6	80.5	−0.89	0.2456	0.0090
F2F vs TA¹	35 to 44	74.1	76.8	−2.68	0.0003	0.0262
F2F vs TA	45 to 49	74.9	77.0	−2.04	0.0551	0.0203
F2F vs TA	50 to 54	75.9	77.2	−1.23	0.1985	0.0125
F2F vs TA	55 to 59	76.0	76.2	−0.28	0.7568	0.0029
F2F vs TA	60 to 64	74.6	75.0	−0.40	0.6623	0.0040
F2F vs TA	65 and over	67.5	65.2	2.31	0.3087	0.0209

¹ Result is statistically significant at (or below) the .05 level.

† Where expected cell counts are small, Chi-square assumptions may not hold; Fisher’s exact test should be used and results interpreted cautiously.

Table D.11 WCA prognosis period by allocated channel

Allocated channel	N	Mean (months)	Standard deviation
F2F	14,665	21.02	6.57
VA	5,210	20.96	6.55
TA	50,039	20.89	6.70

Table D.12 Test of association between WCA prognosis period and allocated channel

Comparison	Test	Mean diff (months)	95% CI	t	Probability-value
F2F vs VA	Satterthwaite	0.06	−0.15 to 0.26	0.5	0.593
F2F vs TA¹	Satterthwaite	0.13	0.01 to 0.25	2.1	0.036

¹ Result is statistically significant at (or below) the .05 level.

Section E – Channel change analyses

Logistic regression has been used to assess which key characteristics (independent variables) are statistically related to claimants changing channels.

The tests were used to assess the following characteristics:

• primary PIP condition
• gender
• age

As the independent variables considered are categorical variables, the regression analysis provides an insight into which categories within these key characteristics are more or less likely to have caused someone to change channel.

When using categorical variables in regression analysis one of the categories needs to be specified as the baseline group. The odds ratio, EXP (β) of the baseline category, is set as 1. The odds ratios of the other categories are then calculated relative to the baseline category. Where the odds ratio is greater than 1, this category is more likely to have stayed in their allocated channel, compared with the baseline category. Alternatively, where the odds ratio is less than 1, this category is less likely to have stayed in their allocated channel, compared with the baseline category.

Although logistic regression can be used to explore associations between variables, it does not necessarily imply causation, and results should be treated as indicative, rather than conclusive.

Assessments from Scotland and Northern Ireland were excluded from the results due to devolution of powers to the local government, in these regions.

Table E.1 Logistic regression analysis of PIP channel change by primary condition

PIP primary condition	Odds ratio	95% Confidence Interval	Probability-value
Anxiety and depression	Reference category	No data	No data
Psychiatric disorders (excluding anxiety and depression)¹	0.771	0.712–0.835	<0.0001
ADHD or ADD	1.031	0.941–1.129	0.5152
Autism¹	0.79	0.719–0.868	<0.0001
Cardiovascular disease	1.022	0.952–1.098	0.547
Endocrine disease¹	1.147	1.051–1.252	0.0021
Gastrointestinal disease¹	1.218	1.117–1.327	<0.0001
Learning disability¹	0.71	0.586–0.861	0.0005
Malignant disease¹	0.725	0.662–0.793	<0.0001
Musculoskeletal disease (general)¹	1.121	1.076–1.168	<0.0001
Musculoskeletal disease (regional)¹	1.188	1.136–1.241	<0.0001
Neurological disease	0.995	0.947–1.045	0.8321
Respiratory disease¹	1.097	1.028–1.171	0.0054
Unknown¹	2.297	2.002–2.635	<0.0001
All other categories¹	1.14	1.082–1.201	<0.0001
Sample size	254,023	No data	No data
Reference category odds ratio set to 1

Odds ratios below 1 indicate lower odds of being assessed in the originally booked channel. ¹ The result is significant at (or below) the .05 level.

Table E.2 Logistic regression analysis of PIP channel change by age

Age	Odds Ratio	95% Confidence Interval	Probability-value
16 to 24	Reference category	No data	No data
25 to 34	0.989	0.937 to 1.043	0.6753
35 to 44	0.951	0.903 to 1.002	0.0602
45 to 54¹	0.933	0.887 to 0.983	0.0085
55 to 64¹	0.891	0.848 to 0.936	<0.0001
65 and over¹	0.853	0.783 to 0.930	0.0003
Sample size	254,024	No data	No data
Reference category odds ratio set to 1

Odds ratios below 1 indicate lower odds of being assessed in the originally booked channel.

¹ The result is significant at (or below) the .05 level.

Table E.3 Logistic regression analysis of PIP channel change by gender

Gender	Odds Ratio	95% Confidence Interval	Probability-value
Male	Reference category	No data	No data
Female	0.986	0.960 to 1.013	0.3166
Sample size	254,024	No data	No data

Reference category odds ratio set to 1.

Odds ratios below 1 indicate lower odds of being assessed in the originally booked channel.