Improving the effectiveness of sickness benefit case management through a public-private partnership? A difference-in-difference analysis in eighteen Danish municipalities

Table 5 Effects of the intervention with one control municipality for each intervention municipality

	Sickness benefit				Self-support
	Effects for all spells		Distinct effects for spells exceeding 52 weeks		Effects for all spells		Distinct effects for spells exceeding 52 weeks		Number of sickness benefit spells in the intervention and control municipalities
	HR^b	95% CI	HR^b	95% CI	HR^b	95% CI	HR^b	95% CI
All municipalities^a	1.03	0.97–1.09	0.99	0.84–1.17	1.00	0.97–1.03	0.98	0.90–1.07	201,583
Assens	1.09 ( p = 0.001)	1.03–1.15	0.98	0.81–1.18	1.08 ( p = 0.006)	1.02–1.15	1.11	0.92–1.34	23,389
Herning	1.00	0.96–1.04	0.75 ( p < 0.001)	0.65–0.86	1.02	0.97–1.06	0.93	0.80–1.07	40,301
Hjørring	1.04	0.94–1.14	1.26	0.89–1.76	0.96	0.87–1.06	0.97	0.71–1.33	13,284
Holbæk	1.02	0.97–1.07	1.30 ( p = 0.006)	1.08–1.56	0.98	0.93–1.04	0.90	0.75–1.07	30,197
Horsens	1.11 ( p < 0.001)	1.07–1.16	1.31 ( p < 0.001)	1.13–1.52	0.98	0.94–1.03	0.99	0.85–1.16	38,107
Kolding	0.99	0.96–1.03	0.97	0.86–1.10	0.99	0.96–1.03	1.02	0.90–1.15	56,305

Estimates are based on Cox proportional hazard rate models with baseline variables: gender, age, citizenship, months of schooling, marital status, pre-school children, previous sick-listing, number and length of hospitalizations, and number of contacts with general practitioners (GPs), medical specialists, physiotherapists and psychologists. All baseline variables are time-invariant
^aStandard errors are clustered by municipality
^bAn HR > 1 indicates a positive intervention effect. Statistically significant coefficients are in bold with p-value in ()

ISSN: 1471-2458