Why some children hospitalized for pneumonia do not consult with a general practitioner before the day of hospitalization

Figures & data

Table 1. Interview questions designed to elicit reasons for not consulting in general practice.

Condition	Question
Child did not see a GP before the index hospital presentation (as indicated in questionnaire)	Did you think about seeing a GP at any point? (Probe: Was there any particular reason why [child] didn't see a GP?)
Child saw a GP before the index hospital presentation (as indicated in questionnaire)	Was that the first time you thought about taking your child to a GP, or did you think about it earlier in the illness? (Probe: Was there any particular reason why [child] didn't see a GP earlier?)

Table 2. Bivariate analysis of GP consultation before the day of hospital presentation with pneumonia.

Variable	Early consulters (n = 94)	Late and non consulters (n = 57)	P-value
Age (years)	3.82 (1.60–6.62)^a	4.19 (1.79–8.75)^a	NS
Male versus female	56/94 (59.6%)	32/57 (56.1%)	NS
White ethnicity versus other ethnicity	87/94 (92.6%)	55/57 (96.5%)	NS
Illness duration before index hospital presentation (days)	6 (4–12)^a	2 (1–5)^a	< 0.001
Presented to hospital during H1N1 flu pandemic (Apr–Dec 2009 inclusive)	59/94 (62.8%)	36/57 (63.2%)	NS
Presented to hospital during winter season (Oct–Mar inclusive)	65/94 (69.1%)	36/57 (63.2%)	NS
Presented to hospital after June 2009 (median month)	41/94 (43.6%)	26/57 (45.6%)	NS
Submitted to hospital by GP referral^g	62/91 (68.1%)	32/55 (58.2%)	NS
Submitted to hospital by self-referral^g	27/91 (29.7%)	23/55 (41.8%)	NS
Took antibiotics before day of hospital presentation	42/94 (44.7%)	3/57 (5.3%)	< 0.001
Obtained advice from an alternative primary care service^b,c	32/76 (42.1%)	8/45 (17.8%)	0.006
Obtained advice from NHS Direct telephone helpline^c	22/76 (28.9%)	5/45 (11.1%)	0.02
Obtained advice from NHS Direct website^c	7/76 (9.2%)	0/45 (0.0%)	0.05
Obtained advice from a pharmacist/chemist^c	10/76 (13.2%)	3/45 (6.7%)	NS
Obtained advice from a complementary/alternative therapist^c	0/76 (0.0%)	1/45 (2.2%)	NS
Uses inhaler for asthma	20/94 (21.3%)	15/57 (26.3%)	NS
Ongoing medical condition apart from asthma^d	14/94 (14.9%)	9/57 (15.8%)	NS
Chickenpox during month prior to illness	3/94 (3.2%)	4/57 (7.0%)	NS
Number of times child saw a GP in the past year for any other illness^e	2 (0–4)^a	2.25 (1–3.25)^a	NS
Carer is lone parent	14/94 (14.9%)	10/57 (17.5%)	NS
Firstborn child of carer	46/94 (48.9%)	23/57 (40.4%)	NS
Only child^f living at home	27/94 (28.7%)	14/57 (24.6%)	NS
Attends school, nursery or day care	75/94 (79.8%)	45/57 (78.9%)	NS
Number of people per bedroom living at home (crowding)	1.33 (1.00–1.33)^a	1.25 (1.00–1.67)^a	NS
Ratio of children^f to adults living at home	1.0 (0.5–1.0)^a	1.0 (1.0–1.5)^a	0.04
Carer/family owns home	79/94 (84.0%)	36/57 (63.2%)	0.004
Daily smoker living at home	27/94 (28.7%)	23/57 (40.4%)	NS
Welsh Index of Multiple Deprivation (WIMD):			0.01
1st quintile	15/93 (16.1%)	14/57 (24.6%)
2nd quintile	15/93 (16.1%)	4/57 (7.0%)
3rd quintile	29/93 (31.2%)	8/57 (14.0%)
4th quintile	18/93 (19.4%)	11/57 (19.3%)
5th quintile	16/93 (17.2%)	20/57 (35.1%)

NS, not significant (P > 0.05).

^aMedian (interquartile range).

^bNHS Direct telephone helpline or website, pharmacist, or complementary/alternative therapist

^c30 participants (19.9%) were excluded from analysis as the questionnaire was completed after discharge from hospital and, therefore, it was not clear whether advice was obtained before or after hospital presentation. Data were assumed to be missing completely at random.

^dCarer responses (n ≥ 2) included eczema (n = 6), food allergies (n = 4) and glue ear (n = 2).

^eIf carer responded with a range instead of a precise value (e.g. ‘3 or 4’ times), the midpoint was used for analysis (e.g. 3.5).

^fAge < 16 years.

^gHospital submission data were missing for five (3.3%) cases (three early consulters and two late/non-consulters). In addition, two early consulters were excluded from both the ‘GP referral’ and ‘self-referral’ categories: one was referred by NHS Direct and one was referred by an unspecified ‘other immediate’ route.

Figure 1. Invitation, recruitment and selection of participants for analysis.

Table 3. Bivariate analysis of indicators of illness severity.

Variable	Early consulters (n = 94)	Late and non consulters (n = 57)	Odds ratio (95% CI) for early consultation	P-value
Tachypnoeic^a at presentation	31/92^c (33.7%)	20/55^c (36.4%)	0.79 (0.36–1.74)	NS
Tachycardic^b at presentation	54/94 (57.4%)	40/57 (70.2%)	0.57 (0.29–1.16)	NS
Temperature > 38.5°C at presentation	20/94 (21.3%)	18 (31.6%)	0.59 (0.28–1.23)	NS
Length of hospital stay (days): median (interquartile range)	2 (1–5) days^c	2 (1–3) days	1.60 (0.81–3.14) for > 2 days	NS
Treated with intravenous antibiotics	53/90^c (58.9%)	37/53^c (69.8%)	0.62 (0.30–1.27)	NS
Discharge diagnosis of empyema	15/94 (16.0%)	6/57 (10.5%)	1.61 (0.59–4.43)	NS

NS, not significant (P > 0.05).

^aDefined as above the normal range for age: > 40 breaths per minute (bpm) for age < 1 year, > 35 bpm for age 1–2 years, > 30 bpm for age 2–5 years, > 25 bpm for age 5–12 years and > 20 bpm for age > 12 years (29).

^bDefined as above the normal range for age: > 160 beats per minute (bpm) for age < 1 year, > 150 bpm for age 1–2 years, > 140 bpm for age 2–5 years, > 120 bpm for age 5–12 years and > 100 bpm for age > 12 years (29).

^cIndicates missing data in medical records (< 5% of subsample).

Table 4. Multivariate analysis of pre-presentation variables associated with early GP consultation (n = 121).

Variable	Crude OR (95% CI)	Adjusted OR (95% CI)	P-value
Illness duration ≥ 3 days^a	7.98 (3.64–17.49)	4.36 (1.67–11.39)	0.003
Took antibiotics before day of hospital presentation	14.54 (4.24–49.82)	10.35 (2.16–49.55)	0.003
Obtained advice from an alternative primary care service^b,c	3.36 (1.38–8.19)	2.20 (0.81–6.01)	0.13
Carer/family owns home	3.07 (1.42–6.64)	3.17 (1.07–9.37)	0.04

OR, odds ratio; CI, confidence interval.

^aIllness duration before hospital presentation was not linear in the logit of GP consultation, therefore, it was converted into a binary variable.

^bNHS Direct telephone helpline or website, pharmacist, or complementary/alternative therapist.

^cThirty participants (19.9%) were excluded from analysis as the questionnaire was completed after discharge from hospital and, therefore, it was not clear whether advice was obtained before or after hospital presentation. Data were assumed to be missing completely at random.

Table 5. Reasons for not seeing a GP before the day of hospital presentation (n = 28).

Reason^a	n	%
• Carer did not consider consulting a GP earlier	11	39.3
• Carer did not think earlier symptoms were serious/unusual	10	35.7
– No further details given	3	10.7
– Child seemed to improve initially	2	7.1
– Presumed asthma	2	7.1
– Thought abdominal pain was due to a pulled muscle	1	3.6
– Child was active/alert	1	3.6
Child’s breathing was normal	1	3.6
– Presumed common self-limiting infection	1	3.6
• GP surgery closed (out of hours)	5	17.9
– Decided to wait until surgery open	4	14.3
– Consulted NHS Direct instead	1	3.6
• Difficulty getting GP appointment	2	7.1
– No emergency appointments available	1	3.6
– Surgery advised carer not to bring child in due to chickenpox	1	3.6
• A&E used as alternative to GP	2	7.1
– Sudden onset of severe symptoms	1	3.6
– Out-of-hours (Thursday afternoon) not covered by any GP service	1	3.6
• Negative preconceptions	2	7.1
– Afraid of wasting the doctor's time	2	7.1
– Anticipated GP would diagnose viral infection	1	3.6
• Unable to travel to GP surgery and GP declined home visit	2	7.1
– Child too ill to move	1	3.6
– Carer had no car	1	3.6

^aCategories are not mutually exclusive; some carers gave more than one reason.

Advanced Life Support Group. The seriously ill child. In: Samuels M, Wieteska S, editors. Advanced paediatric life support: The practical approach. 5th ed. Oxford: Wiley–Blackwell; 2011:58.

 Google Scholar