Soc Psychiatry Psychiatr Epidemiol DOI 10.1007/s00127-015-1041-2

ORIGINAL PAPER

A work time study analysing differences in resource use between psychiatric inpatients Jan Wolff1,2 • Paul McCrone1 • Mathias Berger3 • Anita Patel1 • Christian Ackermann3 Katja Gerhardt3 • Claus Normann3



Received: 7 November 2014 / Accepted: 6 March 2015  Springer-Verlag Berlin Heidelberg 2015

Abstract Purpose Judgement about the adequacy of reimbursement schemes requires an understanding of differences in resource use between patient groups. The aim of this study was to analyse staff time allocation of psychiatrists, psychologists and nurses in inpatient mental health care and to use these data to analyse differences in per diem resource use between patient groups. Methods A self-reporting work-sampling study was carried out at a psychiatric teaching hospital. All of 36 psychiatrists, 23 psychologists and 106 nurses involved in clinical care during the study period participated in a twoweek measurement of their work time. Results A total of 20,380 observations were collected, representing about 10,190 h of work or 6.2 full-timeequivalent years. The average resource use in minutes of staff time per patient day was 227 min, representing 138 € of staff costs. The most resource intensive care was provided at the Psychiatric Intensive Care Unit and for geriatric patients with 334 and 266 min per patient day (192 and 162 €), respectively. The least resource intensive care was provided for patients with substance-related disorders (197 min, 116 €). Substantial shares of clinical work time

& Jan Wolff [email protected] 1

Institute of Psychiatry, Psychology and Neuroscience, Health Service and Population Research, Centre for the Economics of Mental and Physical Health, King’s College London, London, UK

2

Department of Management and Controlling, Medical Centre-University of Freiburg, Freiburg, Germany

3

Department of Psychiatry and Psychotherapy, Medical Centre-University of Freiburg, Freiburg, Germany

were dedicated to tasks without patient contact (58 %). Nursing time was the main driver of total resource use, representing 70 % of staff time and 60 % of costs. Conclusion Presented differences in per diem resource use should inform discussions about the adequacy of reimbursement schemes. Tasks in the absence of the patient, such as documentation and administration, should be reduced to free resources for direct patient care. Keywords Psychiatry  Hospitals  Inpatients  Economics  Costs and cost analysis  Prospective payment systems

Introduction Prospective payment schemes are used or being implemented in Europe and in the USA. Such schemes use classification systems to define differential payment rates for patient groups. These groups should be relatively homogenous in terms of resource use to set an adequate payment for all patients within a class [1]. The implicit assumption underlying the requirement of payment according to patient classification systems is that required resource use systematically differs between groups of patients and that these differences are attributable to measurable patient characteristics. Research in the context of the unsuccessful implementation of Diagnosisrelated Groups for mental health care during the 1980s in the USA found that it was difficult to associate resource use in terms of length of stay to patient characteristics [2]. A more recent systematic review found comparably consistent differences in per diem resource use between patients grouped by diagnoses [3]. However, available studies were relatively old and only one smaller study was carried out in

123

Soc Psychiatry Psychiatr Epidemiol

European settings. Therefore, more recent studies in European hospitals could greatly inform current local developments. Costing of health care interventions requires an adequate measurement of relevant resource use [4]. Staff is the dominant cost factor in inpatient mental health care [5]. Measurement of staff time is usually more complex than other resource use, since ‘‘movements of men […] leave nothing visible or tangible behind them’’ (Taylor [6]). While automation has mainly reduced the need for work time measurement in industrial economics [7], health care processes are less standardised, due to factors such as professional discretion and complexity [8]. Two common methods of measuring staff time use are time-and-motion studies and work sampling [9]. Time-andmotion studies require continuous observation of study subjects. In contrast, work sampling employs snapshots of subjects’ activities at random instants to extrapolate the overall work time distribution [10]. The validity of work sampling results has long been established [11]. Furthermore, it is considered to be more efficient and allows incorporating large numbers of staff [12]. The aim of this study was to analyse the work time allocation of psychiatrists, psychologists and nurses in inpatient care. Furthermore, the aim was to use these data to analyse differences in per diem resource use between patient groups.

Methods A 2-week self-reporting work sampling study was carried out at the psychiatric hospital of the Medical CentreUniversity of Freiburg, Germany. The hospital has eight wards with a total of 120 beds. It provides care to about 1000 inpatient episodes per year. All of 36 psychiatrists, 23 psychologists and 106 nurses involved in clinical care during the study periods in December 2013, May 2014 and March 2014, respectively,

participated in an analysis of their work time. A handheld device specifically for the purpose of work-sampling studies was used, consisting of a pager and documentation forms. The pager gave acoustic signals at random instants during 30-min intervals, upon which participants documented their current place and type of activity. The study included the whole shifts. On call duties were excluded. The number of surveyed types of activities was kept small to reduce complexity. They are summarised in Table 1. First, ‘therapy sessions’ included all planned direct patient contacts with therapeutic purpose and at least 25 min of duration, which had to be coded for reimbursement, including both individual and group therapies. Second, ‘routine care’ included all other direct patient contacts, usually of less than 25 min, such as ward rounds and brief patient interactions. Third, ‘indirect care’ comprised all activities dedicated to a specific patient in the absence of the patient, such as documentation, writing medical letters, care planning and supervision. Fourth, ‘general clinical activities’ comprised tasks that were not related to a specific patient but required for maintaining the inpatient care processes, such as organisational meetings and staff planning. Furthermore, the categories ‘research’, ‘teaching’, ‘training’ and ‘others’ were surveyed to exclude time spent in such activities from resource use analysis. The surveyed places were wards for patients with mainly schizophrenia (13 beds), substance-related disorders (14 beds), geriatric disorders (12 beds), affective disorders (21 and 19 beds), personality disorders (14 beds) and a locked Psychiatric Intensive Care Unit (ICU, 14 beds). At least 50 % of patients admitted to these wards either had a main diagnosis from the same diagnostic group, were of more than 65 years of age (ward for geriatric disorders) or were classified as intensive care patients and had to be treated at the locked ICU, respectively. Furthermore, the outpatient departments, a unit for the diagnostics of sleeping disorders and scientific units were surveyed to exclude time spent at these places from analysis of inpatient resource use.

Table 1 Surveyed types of activities Type

Description

Example

Therapy sessions

Planned direct patient contacts with therapeutic purpose and at least 25 min of duration, which had to be coded for reimbursement

Individual and group therapy

Routine care

Direct patient contacts of usually less than 25 min, which were not coded for reimbursement

Ward rounds and brief patient interaction

Indirect care

Activities dedicated to a specific patient in the absence of the patient

Medical documentation and writing medical letters

General clinical activities

Activities not related to a specific patient but required for maintaining the inpatient care processes

Organisational meetings and staff planning

Research, teaching, training, others

Activities not related to inpatient care, which were surveyed to exclude time from inpatient resource use analysis

writing papers, lecturing and breaks

123

Soc Psychiatry Psychiatr Epidemiol

Participation was anonymous. Staff representatives were involved in conceptualisation of the study and the underlying rationale was clearly described beforehand to reduce observer effects. The study and its methodology were approved by the works council of the Medical CentreUniversity of Freiburg and by the ethics committee of the Albert-Ludwigs-University of Freiburg. Patient age, main diagnosis and length of stay were derived from the patient administration database. Total numbers of observations was used to determine absolute work time, using one observation to represent 30 min. Per diem resource use was calculated by multiplying estimated proportions by total work time per ward and dividing the products by the number of patient days during the study period. Staff time was monetarily valued, including nonwage labour costs, according to the salary and working hours defined in the collective bargaining agreements applicable to university hospitals in Baden-Wuerttemberg, resulting in 0.84 € per minute of psychiatrists’ work, 0.73 € per minute of psychologists’ work and 0.52 € per minute of nurses’ work. Statistical analysis included maximum-likelihood estimations for total proportions of work time and differences between staff groups. Robust standard errors were calculated for confidence intervals and significance tests to adjust for autocorrelation between subsequent observations by clustering each shift. Missing data were addressed by list wise deletion if the number of observations with

missing information was very small, i.e. less than 1.5 %, since negative effects on validity of results were unlikely [13]. Otherwise, data were included in analysis and missing information was reported in tables.

Results A total of 20,380 observations were collected, representing about 10,190 h of work or 6.2 full-time-equivalent years. The work time distribution of psychiatrists and psychologists among types of activities is provided in Table 2. Both dedicated about 61 % of their total work time to specific patients in therapy sessions, routine care or indirect care. About 16 % per cent was required for general clinical care. A remainder of 23 % was spent in activities not mainly dedicated to inpatients, such as research and outpatient care, which were excluded from later analysis of inpatient resource intensity. Differences were found in work time allocation between the two professional groups. Psychologists spent more time in therapy sessions (26 %) than psychiatrists, who in turn dedicated more time to routine care (24 %). In both professions, senior staff required less time for indirect care (17 and 24 %) than their colleagues. Despite overlapping confidence intervals, the difference between psychologists in training and licensed psychologists was approaching conventional levels of significance (p = 0.051).

Table 2 Work time distribution of psychiatrists and psychologists among types of activities Psychiatrists

Psychologists

Resident (n = 922) (%)

Attending physician (n = 796) (%)

All (n = 1718) (%)

In training (n = 397) (%)

Licensed (n = 708) (%)

All (n = 1105) (%)

Therapy sessions

19.36

7.60

13.91

27.45

22.45

25.66

16.69–22.36

5.49–10.42

11.97–16.11

24.77–30.32

17.49–28.31

23.09–28.39

Routine care

23.37

23.99

23.66

8.05

8.45

8.19

21.2–25.71

20.18–28.27

21.52–25.96

5.73–11.19

5.5–12.76

6.29–10.61

29.83

17.15

23.95

29.08

23.71

27.15

27.27–32.52

14.34–20.38

21.84–26.21

25.96–32.41

19.78–28.15

24.66–29.8

16.97

15.70

16.39

17.64

11.85

15.57

14.78–19.42 1.14

12.58–19.43 15.58

14.47–18.5 7.83

14.84–20.85 3.95

8.93–15.56 13.87

13.4–18.02 7.51

Indirect care General clinical Research Teaching Other Total

0.6–2.17

12.14–19.77

5.96–10.22

2.08–7.4

7.73–23.65

4.79–11.58

2.01

3.14

2.53

6.56

12.11

8.55

1.19–3.37

2.02–4.84

1.81–3.54

4.5–9.47

7.78–18.35

6.39–11.35

7.32

16.83

11.73

7.27

7.57

7.38

6.2–8.63

13.39–20.95

9.88–13.87

5.98–8.81

5.68–10.02

6.27–8.65

100.00

100.00

100.00

100.00

100.00

100.00

Bold values indicate maximum-likelihood estimators, range 95 % confidence intervals n number of working hours surveyed

123

Soc Psychiatry Psychiatr Epidemiol Table 3 Work time distribution of nurses among types of activities Assistant (n = 290) (%)

Licensed (n = 2992) (%)

Specialist (n = 1359) (%)

Head nurse (n = 509) (%)

Unknown (n = 862) (%)

All (n = 6012) (%)

Therapy sessions

1.72

8.82

9.13

4.91

7.25

7.99

0.54–5.37

7.7–10.1

7.42–11.17

3.12–7.63

5.23–9.97

7.2–8.87

Routine care

20.00

27.34

26.94

18.45

31.15

26.69

16.81–23.61

25.62–29.13

24.81–29.19

15.19–22.24

28.17–34.3

25.54–27.87

40.86

36.85

35.44

28.66

35.90

35.89

35.3–46.68

35.02–38.7

33–37.97

24.9–32.72

32.54–39.41

34.64–37.17

General clinical

27.59 22.43–33.43

14.99 13.17–17.02

16.49 13.63–19.8

23.65 18.82–29.28

17.63 14.1–21.84

17.05 15.67–18.51

Training

0.52

6.27

6.96

16.98

3.71

6.69

0.07–3.56

4.53–8.61

4.44–10.74

11.5–24.34

1.66–8.11

5.4–8.26

9.31

5.73

5.04

7.36

4.35

5.69

7.57–11.4

5–6.56

4.25–5.97

6.08–8.88

2.9–6.49

5.19–6.24

100.00

100.00

100.00

100.00

100.00

100.00

Indirect care

Other Total

Bold values indicate maximum-likelihood estimators, range 95 % confidence intervals n number of working hours surveyed, unknown missing documentation of seniority level

Table 3 provides the work time distribution of nursing staff among types of activities. Seventy-one per cent of nurses’ total work time was dedicated to specific patients either in therapy sessions, routine care and indirect care. Seventeen per cent was spent with general clinical care. The remainder of 12 % was spent in activities not mainly dedicated to inpatients, such as research and outpatient care, which were excluded from later analysis of inpatient resource intensity. Nurses spent less time in therapy sessions (8 %) and more time in routine care (27 %) than both psychiatrists and psychologists. Similar to the other professions, shares of indirect care of nurses’ total work time monotonically declined with increasing seniority from 41 % in assistants to 29 % in head nurses. Figure 1 shows per diem resource use in minutes (min) of staff time of psychiatrists, psychologists and nurses stratified by wards. Work time of other therapists and oncall duties were not included. The mean daily use of staff time in the complete sample was 227 min per patient. Weighted by average personnel costs, this represented 138 €. The most resource intensive care was provided at the ICU and the ward for geriatric disorders, with 334 and 266 min or 192 and 162 € per patient day, respectively. The least resource intensive care was provided at the ward for substance-related disorders with 197 min or 116 € per patient day. Nursing time was the main driver of total resource use, representing 70 % of staff time and 60 % of costs per patient day in the complete sample. The intensity of care of psychiatrists, psychologists and nurses was unevenly distributed among wards. While care of psychiatrists and nurses was most resource intensive at the

123

ICU and the ward for geriatric disorders, care of psychologists was most intensive at the wards for personality disorders and for schizophrenia patients and least intensive at the ICU. Figure 2a separates per diem resource intensity into types of care including all staff groups. The least resource intensive activity was therapy sessions (30 min, 20 €). The most resource intensive activity was indirect care (87 min, 52 €). These findings were consistent among all wards. Tasks in the absence of the patient, meaning indirect care and general clinical activities together represented more than half (58 %) of total resource use per day (132 min, 79 €). Routine care was most resource intensive at the ICU (104 min, 60 €). Therapy sessions were most resource intensive at the ward for personality disorders and the ward for affective disorders (34 min, 23 €). The time required for general clinical activities was mainly similar between the wards, except of the ICU with 86 min or 47 € per patient day. The unequal shapes of Fig. 2b–d illustrate the different work time allocations of psychiatrists, psychologists and nurses. For instance, the most resource intensive activity of psychiatrists at the ward for personality disorders was therapeutic interventions, in contrast to the ward for geriatric disorders, where they were their least resource intensive type of activity. Psychologists spent most time with therapy sessions and indirect care at all wards, with the later mainly following the former, except in the ward for geriatric patients. The shape of plot 2c for nurses appears very similar to the plot for all staff as a result of the large influence of nursing time on the overall means.

Soc Psychiatry Psychiatr Epidemiol 400

Minutes of staff time per patient day

350

300

250

200

150

100

50

0 ICU

schizophrenia

substance psychiatrists

geriatric psychologists

affective nurses

personality

all

95% CI

Fig. 1 Per diem resource use in minutes of staff time of physicians, nurses and psychologists stratified by wards

Discussion The aim of this study was to analyse the work time allocation of psychiatrists, psychologists and nurses in inpatient care and to use these data to analyse differences in resource use between patient groups. Staff dedicated the main share of their clinical work time to care directly or indirectly for specific patients (80 %). However, substantial shares of clinical work time were required for activities without direct patient contact (58 %). The time spent for indirect care, such as documentation and writing medical letters, decreased with increasing seniority in all staff groups. Psychologists dedicated more time to planned therapeutic sessions. Psychiatrists and nurses required more time for brief interactions within routine care. The most resource intensive care was provided at the ICU and the ward for geriatric patients with 334 and 266 min per patient day, representing per diem costs of 192 and 162 €, respectively. The least resource intensive care was provided at the ward or substance-related disorders with 197 min and 116 € per patient day, representing about one-third of per diem costs at the ICU. The differences in per diem resource use found by this study were similar to the results of previously conducted studies that were included in a recent systematic review [3]. Cromwell et al. [14] found patients with dementia to require the most resource intensive care with a day-mix index (DMI) of 1.21, meaning they required 21 % more

resources per day than the average patient. Patients with substance-related disorders were less resource intensive than average with a DMI of 0.91. Furthermore, also Buckingham et al. [15] and Hirdes et al. [16] found substance-related disorders to be the least resources intensive group of patients with a DMI of 0.78 and 0.80, respectively. However, Cromwell et al. [14], Hirdes et al. [16] and Buckingham et al. [15] consistently found patients with schizophrenia less resource intensive than average, which was only true for psychiatrists and nurses in this study. Strength of this study was the inclusion of all psychiatrists, psychologists and nurses that participated in clinical care during the study period. A measurement of this scale would have been prohibitively costly with other methods, such as constant observation by external observers. On the other hand, the chosen method did not allow measurement of patient-specific resource use. Instead, average resource use was measured at wards mainly dedicated to specific diagnostic groups and average per diem resource use was calculated. Patients might have been different beyond diagnostic grouping, such as different severities of disease, comorbidities or lengths of stay. Moreover, there might have been other factors that have influenced resource use in addition to patient clientele. For instance, Cromwell et al. [14] found per diem resource use to decline by 1.5 % for each increase of 10 % in facility size, likely related to minimum staffing requirements on small units. In this study, most wards were of similar size.

123

Soc Psychiatry Psychiatr Epidemiol ICU

(a) all groups

ICU

(b) psychiatrists

120 100 80

all

schizophrenia

all

schizophrenia

60 40 20 0 substancerelated

personality

affective

ICU

geriatric

personality

therapy

routine

indirect

general

substance-related

affective

ICU

geriatric

14 12 10

all

schizophrenia

all

schizophrenia

8 6 4 2 0 substancerelated

personality

affective

(c) psychologists

personality

substance-related

affective

geriatric

geriatric

(d) nurses

Fig. 2 Per diem resource use in minutes of staff time for different activities of care. Axes in figure segments a–d are scaled differently

The only two wards of substantially different size were those for affective disorders with 19 and 21 beds and these were above average in terms of resource use for psychiatrists and psychologists. However, it cannot be ruled out that the comparably low resource intensity of nursing care at these two wards was related to its large number of beds. Two further limitations should be considered. First, the study period might not have been representative of regular clinical conditions since admitted patients or staff attendance might have been different to usual circumstances. This caveat was addressed by ex ante choosing study periods that were expected to be mainly representative of the annual average, for instance without unusually high absence due to factors such as major conferences. Ex post analyses showed no clinically relevant differences in bed occupancy and number of admission and discharges during the study periods in comparison to the annual averages (not shown in results). Furthermore, staff did not report unusual circumstances upon inquiries.

123

Second, the study was carried out at a single hospital. Therefore, results reflect the care provided at the study site and they potentially incorporate local idiosyncrasies. This raises the issue of generalisability to other hospitals in Germany and to other health care systems. Inpatient care in Germany is delivered within a common framework defined by the Psychiatric Staffing Regulations [17] that provide a common structure of staff. Transferability of results to other health care systems is less clear, since organisational patterns, hierarchy levels and average income might be different. The comparison of this study to international studies has found rather similar results considering the relative differences in per diem resource use between patient groups. However, the exact degree of transferability to other hospital in Germany and to other health care systems remains unclear and requires further elaboration. The results provided by this study are unprecedented in scale and scope and should inform current discussions considering appropriate differences in reimbursement between classes of patients. Furthermore, small proportions

Soc Psychiatry Psychiatr Epidemiol

of total work time were allocated to therapy sessions and large proportions were required for tasks in the absence of the patient. Reimbursement schemes should provide specific funds for therapy sessions. Tasks in the absence of the patient, such as documentation and administration, should be reduced to free resources for direct patient care. Future research should clarify the external validity of the presented results, in particular with respect to other health care systems. Moreover, research should focus on the delineation of patient-specific resource use to relate the results precisely to patient characteristics beyond diagnostic grouping and controlled for confounders. Furthermore, other service-related factors should be investigated, such as the changes of per diem costs during an individual patient’s stay.

References 1. McCrone P (1994) Allocating resources in psychiatric hospitals according to casemix. Psychiatr Bull 18:212–213. doi:10.1192/ pb.18.4.212 2. McCrone P, Phelan M (1994) Diagnosis and length of psychiatric in-patient stay. Psychol Med 24:1025–1030 3. Wolff J, McCrone P, Koeser L, Norman C, Patel A (2015) Cost drivers of inpatient mental health care: a systematic review. Epidemiol Psychiatr Sci 24(1):78–89. doi:10.1017/S20457960 1300067X 4. Drummond MF, Sculpher MJ, Torrance GW et al (2005) Methods for the economic evaluation of health care programmes, 3rd edn. Oxford University Press, Oxford 5. Cromwell J, Maier J, Gage B et al (2004) Characteristics of high staff intensive medicare psychiatric inpatients. Health Care Finance Rev 26:103–117

6. Taylor FW (1911) The principles of scientific management. Harper & Brothers, New York 7. Fehrle M, Michl S, Alte D et al (2013) Time studies in hospitals. Gesundheitso¨konomie Qual 18:23–30. doi:10.1055/s-00321325601 8. Amalberti R, Auroy Y, Berwick D, Barach P (2005) Five system barriers to achieving ultrasafe health care. Ann Intern Med 142:756–764 9. Maynard HB, Zandin KB (2001) Maynard’s industrial engineering handbook, 5th edn. McGraw-Hill, New York 10. Tippett LHC (1935) 3—statistical methods in textile research. Part 3—a snap-reading method of making time studies of machines and operatives in factory surveys. J Text Inst Trans 26:51–70. doi:10.1080/19447023508661637 11. Wirth P, Kahn L, Perkoff GT (1977) Comparability of two methods of time and motion study used in a clinical setting: work sampling and continuous observation. Med Care 15:953–960 12. Wolff J, McCrone P, Auber G, Fiedler P, Patel A, Wetterauer U (2014) Where, when and what? A time study of surgeons’ work in urology. PLoS ONE 9:e92979. doi:10.1371/journal.pone. 0092979 13. Cohen J, Cohen P, West SG (2002) Applied multiple regression/correlation analysis for the behavioral sciences. Lawrence Erlbaum Assoc Inc, Mahwah 14. Cromwell J, Gage B, Drozd E et al (2005) Psychiatric inpatient routine cost analysis—final report. RTI International, Waltham 15. Buckingham B, Burgess P, Solomon S et al (1998) Developing a casemix classification for mental health services: volume 1—main report. Commonwealth Department of Health and Family Services, Canberra 16. Hirdes JP, Fries BE, Botz C et al (2002) The system for classification of in-patient psychiatry (SCIPP): a new case-mix methodology for mental health. interRAI, Ann Arbor 17. Kaltenbach L, Kunze H (2005) Psychiatrie-Personalverordnung: Textausgabe mit Materialien und Erla¨uterungen fu¨r die Praxis, Auflage: 5. Kohlhammer, Stuttgart

123

A work time study analysing differences in resource use between psychiatric inpatients.

Judgement about the adequacy of reimbursement schemes requires an understanding of differences in resource use between patient groups. The aim of this...
499KB Sizes 0 Downloads 6 Views