Differences in COPD-related readmissions to primary and secondary care hospitals

Abstract

Objective. To study differences in readmissions to primary and secondary care hospitals for exacerbations of chronic obstructive pulmonary disease (COPD). Design. A register-based study. Subjects. The data were gathered from the hospital admissions register of the Finnish National Research and Development Centre for Welfare and Health. The data included all acute periods of treatment received by COPD patients aged over 44 years in 1996–2004 who had a principal or subsidiary diagnosis of COPD (ICD 10: J41–J44), respiratory infection (ICD 10: J00–J39, J85–J86) or cardiac insufficiency (ICD 10: I50), followed by an emergency readmission. Treatment had to have taken place in either a primary care hospital or a specialized ward for respiratory diseases or internal medicine in a secondary care hospital. Main outcome measures. The risk of readmission within a week of discharge, analysed by site of care. Results. The risk of readmission within seven days of discharge is 1.74-fold for a patient treated in primary care compared with a patient treated in secondary care. Conclusions. COPD patients discharged from primary care hospitals have a greater risk of readmission, particularly within a week, than those discharged from secondary care. This risk may be attributed to differences in treatment procedures and arrangement of subsequent care. Thus, in the future, more attention should be paid to primary healthcare resources and staff training.

• Acute exacerbation
• chronic obstructive pulmonary disease
• family practice
• general practice
• hospitalization
• readmission

Treatment periods in hospitals and emergency rooms account for 73% of the total costs of managing COPD [1]. Patients are estimated to suffer an average of two exacerbations per year [2], and this disease together with cardiac insufficiency causes the largest number of readmissions to hospitals in Europe and the United States [3]. Repeated referrals to hospital correlate with the degree of severity of the disease [4], quality of life, hospitalization during the previous year, and hypercapnia on discharge [5]. Even severe cases are admittedly not very closely correlated with readmission [6], which tends to reflect above all the quality of the treatment provided and is without doubt a major source of unnecessary expense for the health care system.

The Finnish healthcare system consists of a primary care level comprising about 250 healthcare centres and a secondary care level that comprises 20 hospital districts, including five based on university hospitals, and each run jointly by a number of local authorities. Since healthcare centres usually have inpatient wards run by general practitioners, general practitioners are responsible for a high proportion of the treatment provided for exacerbations of COPD in Finland [7].

Little research has been done on the quality of treatment provided for exacerbations of COPD in primary care hospitals.

• The risk of readmission within seven days of discharge is 1.74-fold for a patient treated in a primary care hospital compared with a patient treated in secondary care.

• The mean time elapsed before readmission is 151.6 days for patients treated in a primary care hospital and 184.9 for those treated in secondary care.

The purpose of this investigation was to assess the quality of treatment for exacerbations of COPD by comparing readmissions to primary and secondary care hospitals.

Material and methods

The data were gathered from the hospital admissions register of the Finnish National Research and Development Centre for Welfare and Health. The data included all periods of treatment received by patients aged over 44 years in 1996–2004 who had a principal or subsidiary diagnosis of COPD (International Classification of Diseases 10: J41–J44). The acute treatment periods that were followed by an emergency readmission of the same patient were selected from the 210 557 periods identified in this way, and the intervals were calculated (from the date of discharge to the date of readmission for COPD, in days). The principal diagnoses in these cases had been either COPD, respiratory infection (ICD 10: J00–J39, J85–J86) or cardiac insufficiency (ICD 10: I50), and treatment had to have taken place in either a primary care hospital or a specialized ward for respiratory diseases or internal medicine in a secondary care hospital. Cases involving transfer of the patient between these types of care were excluded, as were cases with treatment periods longer than 90 days. If admission and discharge took place on the same day, the length of the treatment period was taken to be one day. This procedure yielded a set of 43 288 emergency treatment periods provided for 14 020 patients for which differences in the need for readmission could be examined in relation to the site of care concerned.

The risks of readmission were calculated by age group and site of care (primary care hospital, a ward of respiratory or internal medicine in secondary care) using the free software package R (version 2.2.1). In addition, the first hospitalization periods of the COPD patients treated at a primary care hospital in the interval 1996–2004 (n = 4234) were taken for closer examination of the risk factors leading to readmission within seven days. Odds ratios (OR) with a 95% confidence interval were calculated by binary logistic regression analysis using SPSS 14.0 for Windows (SPSS Inc).

Results

Patients and treatment periods

The patients treated in primary care hospitals, both men and women, were older on average than those in secondary care (Table I), and the mean length of their treatment period was 7.3 days (SD 7.0). The corresponding figure for patients in secondary care was 7.6 days (SD 5.4) when in a ward for respiratory diseases and 6.5 days (SD 5.0) in an internal medicine ward. Altogether 35.9% of the men were treated in primary care hospitals, 41.5% in wards for respiratory diseases, and 22.7% in internal medicine wards in secondary care. The corresponding proportions for women were 29.7%, 46.2%, and 24.0%. The mean interval before readmission was 151.6 days (SD 275.0) in a primary care hospital, 176.5 days (SD 314.6) in a respiratory diseases ward, and 200.5 days (SD 350.1) in an internal medicine ward in a secondary care hospital.

Table I.  Principal parameters of COPD patients for treatment periods beginning with emergency admissions (n = 43 228) by site of care in 1996–2004.

	Primary care hospital	Respiratory diseases ward	Internal medicine ward
Number of admissions by sex
Male	11 679	13 509	7 385
Female	3 186	4 955	2 574
Mean age at admission (years)
Men	74.0	71.0	72.1
Women	73.8	70.0	71.7
Age distribution (%)
< 65 yrs	14.7	23.8	20.3
65–74 yrs	37.2	42.7	41.7
75–84 yrs	39.1	30.4	32.3
≥ 85 yrs	9.0	3.1	5.7
Principal diagnosis (%)
COPD	90.8	86.5	82.6
Respiratory Infection	7.4	12.8	11.7
Cardiac insufficiency	1.7	0.7	5.6
Subsidiary diagnosis given (%)	37.2	45.5	48.0
Cardiovascular disease	14.6	13.0	24.8
Asthma	4.1	7.6	3.6
Cardic insufficiency	3.7	3.2	4.9
Respiratory infection	2.5	4.4	3.1
	1.5	8.1	2.7
Length of treatment period (days)
COPD	7.1	7.5	6.2
Respiratory Infection	8.9	8.1	7.8
Cardiac insufficiency	9.0	7.5	7.9
Readmission (% of treatment periods)
≤ 7 days	14.4	9.5	11.1
8–30 days	26.7	26.7	25.1
31–90 days	23.5	25.6	23.2
≥ 91 days	35.4	38.2	40.6

Risk of readmission

The difference in the risk of readmission diminished once the time interval stretched beyond seven days (Figure 1). In all except the oldest age group, the patients treated in primary care hospitals had a greater risk of readmission within seven days than did those treated in a respiratory diseases ward (Figure 2). Also, the treatment periods on readmission were longer if they commenced within seven days than if they occurred later, with the exception of the patients under 65 years of age treated in primary care and those aged over 84 years treated in an internal medicine ward in secondary care. The former group had a mean length of hospital stay that was 2.8 days shorter than that of the corresponding group treated in a respiratory diseases ward (Table II).

Figure 1.  Relative risk of readmission for COPD patients within 1–7 days, 8–30 days, and 31–90 days of previous discharge in 1996–2004, by site of care (taking the risk of readmission in 1–7 days when treated in a respiratory diseases ward in a secondary care hospital as one).

Figure 2.  Relative risk of readmission in 1–7 days for COPD patients in 1999–2004, by age group and site of care (taking the risk of readmission at age under 65 years when treated in a respiratory diseases ward in a secondary care hospital as one).

Table II.  Mean lengths of treatment periods for COPD followed by readmission in 1999–2004 (n = 43 228) in days, by age groups and site of care.

	Primary care hospital		Respiratory diseases ward		Internal medicine ward
	A	B	A	B	A	B
<65 yrs	4.9	5.6	7.7	7.1	5.8	5.5
65–74 yrs	7.2	6.7	8.4	7.7	6.9	6.3
75–84 yrs	8.3	8.1	8.2	7.7	7.8	6.9
≥85 yrs	9.5	9.1	8.3	7.4	6.9	7.3
Total	7.3	7.3	8.2	7.6	6.9	6.4

A = readmission within 7 days; B = readmission after 7 days.

The mean treatment time for men readmitted within seven days after treatment in a primary care hospital was 7.7 days (SD 7.9) and that for men readmitted later was 7.3 days (SD 6.8); the corresponding figures for women were 8.6 days (SD 6.8) and 8.5 days (SD 7.5). The factors identified as reducing the risk of readmission to a primary care hospital were female sex and institutional or home care on initial discharge (Table III). Men who did not receive home care had a 1.8-fold (95% CI 1.387–2.333) risk of readmission compared with subsequent home care, whereas the corresponding 1.1-fold (95% CI 0.704–1.684) risk for women was not statistically significant.

Table III.  Risk factors for readmission of COPD patients treated in primary care in 1996–2004 who were readmitted within seven days of discharge (n = 4234).

Variable	n	OR	95% CI
Sex
Male	3233	ref
Female	1001	0.790	0.632–0.986
Age group
< 65 yrs	538	ref
65–74 yrs	1506	0.833	0.590–1.176
75–84 yrs	1725	0.792	0.563–1.115
≥ 85 yrs	465	0.826	0.541–1.260
Length of treatment
1–5 days	1975	ref
6–10 days	1427	1.182	0.938–1.489
11–20 days	639	1.077	0.803–1.444
≥ 21 days	193	0.813	0.527–1.253
Time interval
1996–1999	2259	ref
2000–2004	1975	0.977	0.801–1.192
Principal diagnosis
COPD	3725	ref
Cardiac insufficiency	105	1.113	0.556–2.192
Respiratory infection	404	0.847	0.593–1.208
Subsidiary diagnosis
None	2444	ref
Given	1357	1.054	0.842–1.318
Subsequent care
Home care	1,173	ref
Institutionalized	409	1.051	0.753–1.468
At home (no regular services)	2,651	1.566	1.254–1.957

Discussion

The discharge register data can be regarded as exact [8], as concluded by Pajunen in his comparison of this register with that for cardiovascular diseases [9]. The Finnish National COPD programme and those responsible for the Finnish EBM guidelines for COPD have made vast educational efforts to unify the diagnostic bases for the relevant diseases within the whole country [10]. In practice, the Finnish reporting system during the period of concern here was collecting data on what may be termed the “administrative specialty”. This may have influenced the results.

An exacerbation of COPD is defined as an event belonging to the normal course of the disease that is characterized by a change in the basic level of respiratory distress, coughing, and/or sputum that exceeds the normal extent of diurnal variation [11]. The most common reasons for this are respiratory infections or air pollution [12], but in about a third of all serious cases the reason for the exacerbation cannot be ascertained. Cardiac insufficiency can also lead to similar symptoms, and deterioration in this respect can cause a worsening in the condition of a COPD patient.

The risk of readmission was almost twice as great within seven days of discharge from a primary care hospital than from secondary care, but it was not affected by the length of the initial treatment period, the age of the patient or the timing of the treatment in the former case. The risk in general was less pronounced among women than among men.

It is significant that the risk of readmission was greater among the patients treated in primary care, in spite of the fact that they probably represented milder cases of COPD, as can be deduced from the subsidiary diagnoses. However, the Finnish hospital admissions register does not record the severity of the patient's condition. It is possible that the cohort or population treated in primary care may be completely different from the secondary care cohort and that for some reason these patients were no longer sent for secondary care. Perhaps their condition was not expected to improve even with more intensive treatment. On the other hand, the UK National COPD Audit 2003 found that better resources and more organized treatment correlated with mortality and the length of treatment periods, but not with readmissions [13]. In the present material the risk of readmission within seven days among the patients in the age group below 65 years who were treated at primary care hospitals was, at its greatest, almost two and a half times that of patients attended in secondary care, although the differences between the specializations levelled off somewhat from eight days after discharge onwards. A Spanish report has shown that the risk of readmission within a month of treatment for an acute exacerbation of COPD is largely dependent on the severity of the disease, while readmission within a week is dependent on the quality of the treatment [14]. It would seem, therefore, that the treatment provided at primary care hospitals is not of the same standard as that provided by wards of respiratory diseases or internal medicine in secondary care hospitals.

Since the high risk of readmission among patients discharged from primary care hospitals cannot be explained by the length of treatment or the patients’ general state of health, it is possible that the medication given during treatment of the acute phase was not optimal, and that the medication prescribed for regular use was not intensified sufficiently upon discharge. Corticosteroids have been shown to improve pulmonary function during exacerbations of COPD compared with a placebo and to speed up recovery [15]. The treatment recommendations effective in Finland do not depart in any way from those observed worldwide [16], but it is possible that some general practitioners may not follow these recommendations sufficiently closely [2], [17]. It is also possible that they do not give patients as precise instructions as they would receive from a secondary care hospital.

Thiotropium and inhaled steroids with or without a beta-agonist can be used to alleviate the acute phase of COPD [18], [19], and it is notable that consumption of thiotropium in Finland increased by a factor of about 20 over the period 2002–2004 and consumption of a combination of inhaled steroids and bronchodilators doubled [20]. However, it was not possible to study either the medication given for exacerbations or that prescribed for home use on the basis of the present admissions register data, which may be regarded as one limitation of the present study.

General practitioners in primary care hospitals usually know their patients, which may make the threshold for readmitting them lower than it is for specialists, while the effect of home care in reducing readmissions may be attributed to the close supervision and support provided and prompt intervention in the event of an exacerbation, often without the need for readmission to hospital. Providing patients with effective instruction and guidance can affect readmissions [21], although conflicting results have been reported [22].

It may thus be concluded that COPD patients discharged from a primary care hospital have a higher risk of readmission, particularly within one week, than do those leaving a secondary care hospital ward for respiratory diseases or internal medicine, and that this risk may perhaps be attributable to differences in treatment routines, resources, staff capabilities and arrangement of subsequent care. However, since it is reasonable to keep patients in expensive secondary hospital care for the minimum length of time, it is clear that more emphasis should be placed on treatment of exacerbations of COPD in training provided for primary care staff.