Health indicator recording in UK primary care electronic health records: key implications for handling missing data

Figures & data

Figure 1 Number of records of each health indicator per 100 person-years by sex and age (in years).

Figure 2 Percentage of individuals with a record of each health indicator in the 2000 (purple), 2005 (teal), and 2010 (orange) registration cohorts by calendar year.

Note: The 2000, 2005, and 2010 registration cohorts included individuals who were newly registered with their general practices in 2000, 2005, and 2010, respectively.

Table 1 Number of individuals, median age at registration, and sex distribution among those who were newly registered with general practices in 2000, 2005, and 2010

Year of registration	Number of practices	Number of individuals	Median (Q1–Q3a) age at registration in years	Sex, n (%)
				Male	Female
2000	635	180,871	35 (27–50)	86,179 (48)	94,692 (52)
2005	640	215,609	34 (26–48)	102,367 (47)	113,242 (53)
2010	607	195,491	34 (26–47)	91,970 (47)	103,521 (53)

Note:

a Q1, Q3: first and third quartiles, respectively.

Figure 3 Percentage of individuals with a record of each health indicator in the 2000 (purple), 2005 (teal), and 2010 (orange) registration cohorts by calendar year and disease status.

Notes: (A) Diabetes, (B) myocardial infarction, and (C) stroke. The 2000, 2005, and 2010 registration cohorts included individuals who were newly registered with their general practices in 2000, 2005, and 2010, respectively.

Figure 4 Time (in years) from practice registration to having the first record of each health indicator; and time (in years) at which 1) 50% of the individuals have had their first height, weight, SBP, or alcohol consumption record; 2) 25% of the individuals have had their first total cholesterol record; and 3) 75% of the individuals have had their first smoking status record.

Table 2 Associations of the mean weight measurements and the probability of having weight recorded with sex, age group, social deprivation, and indicators of chronic diseases among individuals who were actively registered in 2010

Variables	Differences in the mean weight measurements (n=1,104,221)			Differences in the probability of having weight recorded (n=3,583,437)
	Difference in mean (kg)a	95% CI	Pb	ORc	95% CI	Pb
Sex			<0.001			<0.001
Men	Base level			1.00	1.55–1.57
Women	−13.45	−13.52 to −13.39		1.56
Age group			<0.001			<0.001
18–24	Base level			1.00
25–29	3.45	3.28–3.61		1.02	1.01–1.04
30–34	5.45	5.29–5.62		0.96	0.94–0.97
35–39	7.65	7.49–7.82		0.84	0.83–0.85
40–44	9.08	8.92–9.25		0.83	0.82–0.84
45–49	9.45	9.29–9.61		0.88	0.87–0.89
50–54	9.30	9.13–9.46		0.97	0.96–0.98
55–59	8.23	8.07–8.39		1.09	1.08–1.10
60–64	6.94	6.78–7.09		1.28	1.27–1.30
65–69	4.85	4.69–5.01		1.57	1.55–1.59
70–74	2.63	2.47–2.80		1.77	1.75–1.79
75–79	−0.20	−0.37 to −0.03		1.77	1.75–1.79
80–84	−3.80	−3.99 to −3.61		1.50	1.48–1.53
85–89	−7.70	−7.93 to −7.47		1.13	1.11–1.15
90–94	−10.7	−11.06 to −10.34		0.78	0.76–0.80
95–99	−14.4	−15.15 to −13.65		0.52	0.50–0.55
Townsend score			<0.001			<0.001
Quintile 1 (least deprived)	Base level			1.00
Quintile 2	0.48	0.39–0.58		1.08	1.08–1.09
Quintile 3	0.81	0.71–0.91		1.17	1.17–1.18
Quintile 4	0.92	0.83–1.02		1.25	1.24–1.26
Quintile 5 (most deprived)	0.23	0.12–0.34		1.43	1.42–1.44
Indicators of diseases
Myocardial infarction	−0.19	−0.34 to –0.04	0.015	2.18	2.15–2.21	<0.001
Stroke	−0.75	−0.89 to –0.61	<0.001	1.38	1.37–1.40	<0.001
Diabetes	7.08	7.01–7.15	<0.001	2.53	2.52–2.55	<0.001

Notes:

a Differences in the mean weight measurements (in kg) from a multivariable linear regression model, conditional on sex, age group, social deprivation, and indicators of chronic diseases.

b P-values from joint Wald tests.

c ORs of having a weight measurement recorded from a multivariable logistic regression model, conditional on sex, age group, social deprivation, and indicators of chronic diseases.