Table 2. Sample moments by group

	Named GP (n = 23 762)			No named GP (n = 42 301)
					A. Before balancing			B. After balancing
Variable	Mean	Variance	Skewness	Mean	Variance	Skewness	Mean	Variance	Skewness
Age, years	47.4	4422	26	44.2	5468	24.1	47.4	4424	26.1
Education	0.889	0.605	0.194	0.939	0.65	0.111	0.889	0.605	0.194
Disp. inc. (decile)	5.64	8.17	-0.0662	5.36	8.86	0.0283	5.64	8.17	-0.0604
Born in Sweden or Nordic region	0.883	0.103	-2.39	0.849	0.128	-1.94	0.883	0.103	-2.39
Mover^a	0.0615	0.0577	3.65	0.111	0.0985	2.48	0.0615	0.0577	3.65
Start date (reg.)	17 835	858 335	0.0377	18 432	323 954	0.117	17 835	858 361	-0.07
Duration (reg.)	1298	611 174	1.31	1038	288 864	0.958	1298	611 158	1.4
Registered t-2	0.866	0.116	-2.15	0.817	0.15	-1.64	0.866	0.116	-2.15
Registered t-3	0.72	0.201	-0.983	0.679	0.218	-0.765	0.72	0.201	-0.983
Switched PCC (2–4 years pre)	0.622	0.235	-0.505	0.481	0.25	0.0744	0.622	0.235	-0.504
Switched GP (2–4 years pre)	0.703	0.209	-0.89	0.543	0.248	-0.173	0.703	0.209	-0.89
First diagnosis of depression	0.338	0.224	0.683	0.373	0.234	0.523	0.338	0.224	0.683
First diagnosis of hypertension	0.318	0.217	0.783	0.28	0.202	0.978	0.318	0.217	0.783
First diagnosis of ischaemic heart disease	0.0529	0.0501	3.99	0.0419	0.0402	4.57	0.0529	0.0501	4
Index year	2012	4.42	0.0325	2013	2.59	-0.333	2012	4.42	0.0972
Any GP visit (3–4 years pre)	0.522	0.25	-0.0861	0.495	0.25	0.0201	0.521	0.25	-0.086
Number of GP visits (3–4 years pre)	1.33	3.9	2.87	1.23	3.56	2.68	1.33	3.9	2.91
Number of nurse visits (3–4 years pre)	1.3	11.5	7.82	1.29	10.9	7.05	1.3	11.5	8.91
Any ED visit (3–4 years pre)	0.131	0.114	2.19	0.114	0.101	2.44	0.131	0.114	2.19
Any GP visit (1–2 years pre)	0.867	0.115	-2.16	0.884	0.102	-2.4	0.867	0.115	-2.16
Number of GP visits (1–2 years pre)	2.76	7.21	2.31	2.83	7.11	2.16	2.76	7.21	2.36
Number of nurse visits (1–2 years pre)	2.18	18.2	6.44	2.67	27.2	4.11	2.18	18.2	4.79

The table shows the first three sample moments — the mean, the variance, and the skewness — for the background characteristics which were used to obtain entropy weights. For the comparison group with no named GP, A shows the moments as observed in the data, and B shows the moments after using the entropy weights. The start date and duration variables refer to the PCC registration that was ongoing at the index date. The dummy variables for being registered in t-2 and t-3 refer to the second and third year before the index date (by definition, everyone was registered the year before the index date). The dummy variables for having switched PCC or GP indicate whether such a switch occurred during the 2–4 years before the index date. The prior care use variables at the bottom of the table are calculated over the 3–4 and 1–2 years before the index date, respectively.
^aAn individual who recently moved from another municipality.
ED = emergency department. Disp. inc = disposable income. PCC = primary care centre. Pre = period before the index date. Reg = registration.