While we struggled with these diffe

While we struggled with these differentiations, we were also deeply involved in data analysis for two conceptual papers for presentation and publication to research audiences in social science and nursing: one on the social psychology of managing ordinary ailments (Olesen, Schatzman, Droes, Hatton, Chico and Chesla 1985) and another on the surgical construction of reality (Hatton and Droes 1987). These papers did not focus on comparison of the HMO and non-HMO samples, but, rather, looked at specific conceptual issues in all data collected to date, for instance, the social psychology of body and self in the mundane ailment and in the more dramatic case of the ostomy. These exercises forced us to clarify these elements in our data in order to make them comprehensible to other audiences and to lift our concepts to the more theoretical levels expected of social science research reports. We returned to the work of comparing the samples for our HMO study sponsor in late winter and early spring of 1986.
By early summer of 1986 this analysis began to produce detailed differences between the samples, namely that the HMO members had a calculating approach to care and services, whereas the non-HMOs’ use appeared haphazard at best. A team analytic memo of 2 June shows that our characterization of the sporadic and disjointed non-HMO approaches had become more abstract:
the non-HMO search for health care is episodic, arising when need does and involves brokering alternatives (acupuncture, etc.) for care. Their framework of relating to care givers is disjunctive rather than adjunctive.
More fully defined, these dimensions were: (1) adjunctive versus disjunctive use, referring to seeking out and co-operating with a health care professional versus not discussing ailments with providers and failing to follow orders; (2) episodic versus periodic, referring to incidental or sporadic use of services versus regular use for examinations, ailments, etc. These distinctions emerged again in the team meeting of 5 June 1986:
While the HMO members assess efficacy of self-care efforts and the capacity of medical science to determine whether they will use professional services, they seek attention when self-care no longer handles the situation or the situation is unfamiliar…the HMO structure which provides access to low cost services supports such calculation. With the non-HMOs one does not get the calculation or assessment that HMOs report…. In the non- HMOs’ episodic approach to seeking services, there is no support for this type of calculation. Could it be that the HMO ideology actually teaches or supports self-care? Let’s check this.
This last question would become a probe for further analysis. We also at this time, as the notes show, detected further differences, responsibility in caring for self, a change from the interpretation of a year before noted above, and abruptness in seeking formal care. This prompted additional re-analyses of previously formulated dimensions:
While the HMOs frequently monitored ailments with considerable sophistication, the non-HMOs do not report in detail such behaviors. Rather, one gets a sense of an abrupt decision to seek care…. The non-HMOs lack the considered judgements regarding health that the HMOs demonstrated. There is cruder management of health, and in particular, self-care…. Unlike the HMO members’ reports which show responsibility in managing one’s illness with self-care prior to seeking professional help, this is not as evident in the non-HMO accounts. We need to re-read their accounts for the dimensions of severity, persistence and consequentiality which we have seen to be related to self-care and seeking help.

Analysis blinded, analysis liberated
Late in the summer of 1986 we were floundering. Self-care and use of services were linked, but within our non-HMO sample we could not make sense of the patterns of those who did and did not have insurance. We could not seem to find a way in which we could compare these subgroups with the HMO members in our first analysis.
Later it became clear that our emphasis on the link between the type of coverage or insurance the non-HMO respondents did or did not have and their use of services blinded our analysis. This blinkering was probably a consequence of our focus on the study question (How are services used by those with and without insurance?). It may also have reflected the tenacity of our earlier analyses which described episodic and disjunctive dimensions of use and implicitly linked these to lack of insurance.
Droes’s methodological memo of 15 August 1986 provided a breakthrough. Rather than looking at the adjunctive-disjunctive and episodic-periodic dimensions of the non-HMOs’ use as related to insurance status, she put aside respondents’ insurance status. She then re-categorized all non-HMO respondents using only two criteria: a recent visit to a health provider and health providers’ services being used by respondents adjunctively, periodically, etc. She then intersected these dimensions so that the cases, without regard to insurance status, could be deftly and parsimoniously located within the four quadrants created by the dimensions intersecting one another (see Figure 6.1).
This broke our stalemate, for it became clear that there were diverse patterns. In the episodic-disjunctive quadrant (upper right) we found,

not surprisingly, a predominance of our non-insured respondents, but we also found those with fee-for-service insurance in this quadrant. Conversely, in the adjunctive-episodic quadrant, the upper left, we saw both non-insured and fee-for-service respondents. Thus, some individuals who were insured acted disjunctively and episodically, while those without insurance and who, for instance, used emergency rooms, behaved in an adjunctive and periodic manner.
To break our analytic stalemate, it was necessary to depart from our solidly entrenched thinking on insurance and literally to reshuffle the analytic process. This breakthrough led to reaffirmation of a higher-order set of categories based on respondent reports, e.g. adjunctive-disjunctive, periodic-episodic. These could be used more productively to characterize respondents rather than the lower-order attribute of the presence or absence or type of insurance. That had not only been unproductive but, as we have shown, blinding.

Analysis of clusters
It still remained to link self-care with the intersecting dimensions of adjunctiveness and periodicity noted above. Here we each took clusters or themes and examined cases in the quadrants to discern differences or similarities. Finding such differences was to prove more elusive than we anticipated, as is suggested in Olesen’s memo of 22 August 1986: Nellie’s memo [Droes, mentioned above] of August 15 led me to resort all the cases into the four quadrants of the use table [Figure 6.1] and read all cases in each quadrant for trust in the body simply to see if there were any discernible differences. No distinctive themes emerged that can be tied to particular groups; rather some of our familiar themes, e.g. relationship of trust in self to trust in body appear in all groups. It is more likely that trust and styles and types of self-care are related in a complex way with the patterns of provider use that we have identified. We will have to examine this carefully when we integrate our analysis for the final report. We then met again as a team, argued and debated over our findings in this thematic analysis. Where disagreements arose, individuals had to sustain their position or criticisms with both data and logic. Where agreements occurred, sharp questioning forced those agreeing to document their agreement. Each team member then wrote a section for the final report on his or her themes.
Final comparisons
Our final step was to compare the HMO and non-HMO samples along the five questions set out at the beginning of the study. Here is our answer to the key question of differences in self-care practices (Olesen et al. 1987): We find no striking differences. Both samples have well-tested strategies for managing mundane ailments, using palliative remedies, consulting others (not providers), evaluating the efficacy of their efforts, using tolerance of symptoms, limits to their own expertise and anticipated consequences as guides. The difference between the samples seems to lie in the departure from self-care strategies to use of services, for the HMO members moved smoothly to use HMO services, whereas the non-HMOs used services impetuously and in some instances after considerable delay. On the basis of this very small sample, we conclude that HMO members linked their self-care practices to their HMO membership, albeit very conservatively, for we were impressed with the extent to which they first relied on self-care practices. The non-HMOs, both those with and without insurance, did not seem to tie self-care to provider services.

CONCLUSION
As is true in all qualitative work, our goal was to raise the respondents’ comments, whilst respecting the integrity of those comments, to a higher order of abstraction or generality that reflected the patterning seen in the comments. In part we could realize this because of our long mutual immersion in the study of self-care. That immersion honed our analytic skills, and also provided a safeguard against any team member, faculty or not, dominating the analysis or manipulating agreement; others simply knew too much and too well the subtleties of self-care ethnography readily to agree even on non-problematic categories or themes.
What did we learn from our collective experience? The enduring necessity to be unremittingly and relentlessly reflexive was borne in on us time and again, as we stumbled through some parts of our analysis, but anticipated others. Equally important, we thought, was the capacity to be flexible, to look around the corners and beyond some of our dimensions and conceptions

Hasil (Bahasa Indonesia) 1:

While we struggled with these differentiations, we were also deeply involved in data analysis for two conceptual papers for presentation and publication to research audiences in social science and nursing: one on the social psychology of managing ordinary ailments (Olesen, Schatzman, Droes, Hatton, Chico and Chesla 1985) and another on the surgical construction of reality (Hatton and Droes 1987). These papers did not focus on comparison of the HMO and non-HMO samples, but, rather, looked at specific conceptual issues in all data collected to date, for instance, the social psychology of body and self in the mundane ailment and in the more dramatic case of the ostomy. These exercises forced us to clarify these elements in our data in order to make them comprehensible to other audiences and to lift our concepts to the more theoretical levels expected of social science research reports. We returned to the work of comparing the samples for our HMO study sponsor in late winter and early spring of 1986. By early summer of 1986 this analysis began to produce detailed differences between the samples, namely that the HMO members had a calculating approach to care and services, whereas the non-HMOs’ use appeared haphazard at best. A team analytic memo of 2 June shows that our characterization of the sporadic and disjointed non-HMO approaches had become more abstract: the non-HMO search for health care is episodic, arising when need does and involves brokering alternatives (acupuncture, etc.) for care. Their framework of relating to care givers is disjunctive rather than adjunctive. More fully defined, these dimensions were: (1) adjunctive versus disjunctive use, referring to seeking out and co-operating with a health care professional versus not discussing ailments with providers and failing to follow orders; (2) episodic versus periodic, referring to incidental or sporadic use of services versus regular use for examinations, ailments, etc. These distinctions emerged again in the team meeting of 5 June 1986:While the HMO members assess efficacy of self-care efforts and the capacity of medical science to determine whether they will use professional services, they seek attention when self-care no longer handles the situation or the situation is unfamiliar…the HMO structure which provides access to low cost services supports such calculation. With the non-HMOs one does not get the calculation or assessment that HMOs report…. In the non- HMOs’ episodic approach to seeking services, there is no support for this type of calculation. Could it be that the HMO ideology actually teaches or supports self-care? Let’s check this.Pertanyaan terakhir ini akan menjadi probe untuk analisa lebih lanjut. Kami juga pada saat ini, sebagai acara catatan, terdeteksi lebih lanjut perbedaan, tanggung jawab dalam merawat diri sendiri, perubahan dari interpretasi dari setahun sebelum dicatat di atas, dan kekasaran dalam mencari perawatan formal. Ini mendorong tambahan kembali analisis sebelumnya diformulasikan dimensi:Sementara penyakit HMO sering dipantau dengan kecanggihan besar, bebas-HMO tidak melaporkan secara terperinci perilaku seperti itu. Sebaliknya, salah satu mendapat sebuah rasa dari keputusan mendadak untuk mencari perawatan... Bebas-HMO kekurangan penilaian dianggap mengenai kesehatan yang ditunjukkan HMO. Ada cruder manajemen kesehatan dan perawatan diri tertentu... Tidak seperti anggota HMO laporan yang menunjukkan tanggung jawab dalam mengelola penyakit dengan perawatan sebelum mencari bantuan profesional, hal ini tidak seperti terlihat dalam account bebas-HMO. Kita perlu untuk membaca kembali akun mereka untuk dimensi keparahan, ketekunan dan consequentiality yang telah kita lihat berkaitan dengan perawatan diri dan mencari bantuan.Analisis buta, analisis dibebaskanDi akhir musim panas tahun 1986, kami telah menggelepar. Perawatan diri dan penggunaan layanan terkait, tetapi dalam sampel bebas-HMO kami kita tidak bisa membuat rasa pola mereka yang melakukannya dan tidak memiliki asuransi. Kita tidak bisa tampaknya menemukan cara di mana kita dapat membandingkan subkumpulan ini dengan anggota HMO dalam analisis pertama kami.Later it became clear that our emphasis on the link between the type of coverage or insurance the non-HMO respondents did or did not have and their use of services blinded our analysis. This blinkering was probably a consequence of our focus on the study question (How are services used by those with and without insurance?). It may also have reflected the tenacity of our earlier analyses which described episodic and disjunctive dimensions of use and implicitly linked these to lack of insurance.Droes’s methodological memo of 15 August 1986 provided a breakthrough. Rather than looking at the adjunctive-disjunctive and episodic-periodic dimensions of the non-HMOs’ use as related to insurance status, she put aside respondents’ insurance status. She then re-categorized all non-HMO respondents using only two criteria: a recent visit to a health provider and health providers’ services being used by respondents adjunctively, periodically, etc. She then intersected these dimensions so that the cases, without regard to insurance status, could be deftly and parsimoniously located within the four quadrants created by the dimensions intersecting one another (see Figure 6.1).This broke our stalemate, for it became clear that there were diverse patterns. In the episodic-disjunctive quadrant (upper right) we found, not surprisingly, a predominance of our non-insured respondents, but we also found those with fee-for-service insurance in this quadrant. Conversely, in the adjunctive-episodic quadrant, the upper left, we saw both non-insured and fee-for-service respondents. Thus, some individuals who were insured acted disjunctively and episodically, while those without insurance and who, for instance, used emergency rooms, behaved in an adjunctive and periodic manner.To break our analytic stalemate, it was necessary to depart from our solidly entrenched thinking on insurance and literally to reshuffle the analytic process. This breakthrough led to reaffirmation of a higher-order set of categories based on respondent reports, e.g. adjunctive-disjunctive, periodic-episodic. These could be used more productively to characterize respondents rather than the lower-order attribute of the presence or absence or type of insurance. That had not only been unproductive but, as we have shown, blinding.Analysis of clustersIt still remained to link self-care with the intersecting dimensions of adjunctiveness and periodicity noted above. Here we each took clusters or themes and examined cases in the quadrants to discern differences or similarities. Finding such differences was to prove more elusive than we anticipated, as is suggested in Olesen’s memo of 22 August 1986: Nellie’s memo [Droes, mentioned above] of August 15 led me to resort all the cases into the four quadrants of the use table [Figure 6.1] and read all cases in each quadrant for trust in the body simply to see if there were any discernible differences. No distinctive themes emerged that can be tied to particular groups; rather some of our familiar themes, e.g. relationship of trust in self to trust in body appear in all groups. It is more likely that trust and styles and types of self-care are related in a complex way with the patterns of provider use that we have identified. We will have to examine this carefully when we integrate our analysis for the final report. We then met again as a team, argued and debated over our findings in this thematic analysis. Where disagreements arose, individuals had to sustain their position or criticisms with both data and logic. Where agreements occurred, sharp questioning forced those agreeing to document their agreement. Each team member then wrote a section for the final report on his or her themes.Final comparisonsOur final step was to compare the HMO and non-HMO samples along the five questions set out at the beginning of the study. Here is our answer to the key question of differences in self-care practices (Olesen et al. 1987): We find no striking differences. Both samples have well-tested strategies for managing mundane ailments, using palliative remedies, consulting others (not providers), evaluating the efficacy of their efforts, using tolerance of symptoms, limits to their own expertise and anticipated consequences as guides. The difference between the samples seems to lie in the departure from self-care strategies to use of services, for the HMO members moved smoothly to use HMO services, whereas the non-HMOs used services impetuously and in some instances after considerable delay. On the basis of this very small sample, we conclude that HMO members linked their self-care practices to their HMO membership, albeit very conservatively, for we were impressed with the extent to which they first relied on self-care practices. The non-HMOs, both those with and without insurance, did not seem to tie self-care to provider services.CONCLUSIONAs is true in all qualitative work, our goal was to raise the respondents’ comments, whilst respecting the integrity of those comments, to a higher order of abstraction or generality that reflected the patterning seen in the comments. In part we could realize this because of our long mutual immersion in the study of self-care. That immersion honed our analytic skills, and also provided a safeguard against any team member, faculty or not, dominating the analysis or manipulating agreement; others simply knew too much and too well the subtleties of self-care ethnography readily to agree even on non-problematic categories or themes.What did we learn from our collective experience? The enduring necessity to be unremittingly and relentlessly reflexive was borne in on us time and again, as we stumbled through some parts of our analysis, but anticipated others. Equally important, we thought, was the capacity to be flexible, to look around the corners and beyond some of our dimensions and conceptions